19/10/2011

* GIS FAQ


A very difficult thing to do is to explain GIS to people whom are not technically inclined simply because GIS is a very technical subject. (In this respect, I did not expect and found it most interesting that "GIS FAQ" is one of my most popular articles) To make matters worse, digital subjects can be very fuzzy. To make matters really bad is if that person(s) whom needs explanation is your boss or top management who know nothing or little about GIS yet have too much ego to admit it (Funny thing, seems to be a norm everywhere). To explain a digital subject in an analogue way that people can quickly and clearly comprehend has always been my greatest challenge and I admit is not easy in fact in general I would say difficult. For example mention the word "Vector" and it will mean different things as well as seen from different visual angles by different people, even for me before I got to know about GIS in detail, Vector is the name of a super sportscar :-D. Thus, I thought it interesting to come up with a GIS FAQ that even a layman could understand with QGIS as the tool so that when we talk about GIS, we are on the same frequency and meaningful to both of us in some way or the other or else it will be nodding all the way but nothing absorbed. I admit my explanation may not be so good nor accurate as compared to academicians or better-off experienced GIS users but for the time being I just wanted to do something and list the questions and answers cramping my mind especially the ones I could not answer well. New queries and answers will be added and old ones amended over time (meaning this page is rather dynamic and the most dynamic in this blog) when I find a better crispier example I am happier to live with. Still, I want like to start somewhere on this never-ending draft and as I babble away but if I do leave you happy then I have succeeded in this technological transfer because that is what Open Source  which I grew to believe is all about ...sharing. By the way, before I say here goes, unlike GIS FAQ, QGIS FAQ deals specifically with QGIS issues which you may like to venture later on:

What is GIS?
A digital age technological system about information on geographical features covering spatial and non-spatial data:
  • "Geographical" because it primarily deals with geographical features;
  • "Information" because it can churn out information from data;
  • "System" because it consist of many inter-related components (that do many things: retrieve, visualize, edit, store, query and analyse GIS data).
Although it fundamentally deals with mapping (in the first stage) and analysis (in the second stage), GIS is more prized for its analytical ability to extract information from geographical data. This is where the fun is. Things get interesting when your data (be it topology and attribute) is "clean" which is not easy to do because they usually involve a voluminous amount of data.

What is web-GIS?
Yes I know but this has nothing to do with GIS for spiders. Firstly, some people think or believe web mapping is web-GIS. Then there are those who feel that web-GIS is being able to do GIS over the Internet but with the help of a compatible desktop GIS application to link to, which is the general acceptance of the term 'web-GIS'. But for me,  I am of the opinion that it is a GIS that can do all the above yet over the Internet and on condition, with only an internet browser. Seriously, if a person says otherwise, I say someone is bluffing you into making you believe whatever he or she says really means web-GIS. To me, you do not need a GIS application in your desktop computer for web-GIS after all if you have an internet email account you do not need a email application installed in your computer, do you? If you play internet games you don't need anything installed in the desktop computer, do you? (short of the multimedia drivers). When you use online banking facilities, you don't need to install a specific on-line banking application, do you? You have be clear between web-mapping and web-GIS. There are a lot of con people out there eager to take you for a ride. Take my advise. Do not immediately accept any interpretation immediately including mine. Use your simple logic and common sense. Has GIS come to this level of web-GIS as I see it? You tell me.

Is there any future in Web-GIS?
Before GIS Mobile came into the scene, I would say there was a rush for the bandwagon on anything that even smells of Web-GIS. It appeared to be the glamorous side of GIS and many a superduper senior manager felt their agency must have one let alone whether they really knew what they were taking about when they mention the word "Web-GIS". Now that GIS Mobile is here to stay and looking optimistically the stable version of QGIS Mobile should not be too far away, it should not be long before the functions of a desktop-GIS can be replicated by a GIS Mobile far easier than it is for a web-GIS to deliver, couple to that, there is the advantage of mobility which web-GIS just cannot meet. With the current price of a tablet being corporately reasonable, I could be wrong but I forsee the rise of GIS Mobile and the fall of the Web-GIS empire as people see it and that is where I would put my money on.

What is 3D GIS?
This is another fancy terminology where people have their own interpretation. People were initially brainwashed into thinking that 2 1/2D is 3D and subsequently in the GIS industry associated 3D GIS would take up that perspective. However, have you seen any 3D movies lately where currently you need need 3D spectacles because the movie industry has not yet come up with a 3D screen (The movie industry is currently trying  hard to come up with 3D movies that do not rely on 3D spectacles)? Well, that is my interpretation of "3D" in that there is the element of solid form and depth of field, anything less than that, I do not accept. So if it is a GIS that can do all the above as I defined earlier in "What is GIS?" and on condition in 3D as I see it then I say that GIS software can do 3D GIS. This is my stand.

What is 'spatial data'?
  • Data that has a spatial context. 
This means the locational factor of a geographical feature is taken into account recognised as the meeting point of the latitude and longitude for that feature usually in terms of degrees. Examples of a spatial data is a fire hydrant (point), road (line) and lot parcel(polygon). If altitude (z) data is available, a computed 3D image and subsequent 3D information can be acquired with the help of a 3D module.

What is 'non-spatial data'?
  • Data that has no spatial context.
Since GIS deals with geographical features, the non-spatial contexts refer to attributes of a geographical feature. It can be kept in a spreadsheet or database and can linked to the spatial data in the form of an attribute table or dynamically linked to the spatial data by a hyperlink. Examples of a non-spatial data e.g. lot parcel could include landowner, land ownership status, a photograph or even a video clip of that lot parcel.


What is the difference between data, metadata and information?
  • "Data" is the plural for datum which is basic information about an item. An example of landuse data is  specific information on various types of activities on the land e.g. residential, commerical and industrial.
  • "Metadata" is data about that data or how a particular data shall be structured so that the computer understands what you are talking about e.g. when you say 'name', the computer will wants to know (it has not learnt how to 'ask' yet) the number of characters to capture names say '20' anything more, it will not recognize as well as the form names shall take that either alphanumeric (string) or just numeric (integer).
  • "Information" is specific issues analysed from that data e.g. 20% of the existing bungalows in a study area are located on leasehold land.
The sad thing is that many people, top management included, are confused what type of data they should collect. They will say we want all type of data but I bet you, ask them why and how it will be of use to them and what is the GIS analytical relationship between data K and data Q or what data is required to acquire a specific information and how do you go about analytically getting it, I am quite sure they do not know how to. Data collection is the most expensive part of a study so it is wiser to understand in the first place, which information is only required to obtain a certain  information. No use and a waste of time and money to collect non-relevant data. On the issue of metadata, I was once asked and replied I did not know how to input the parameter for a ridge for a certain slope analysis then got ridiculed because to a researcher, a 'ridge' was so obvious as if I was trying to make things difficult. Then I replied "I can give you a ridge with a 60 degree slope and a 10,000 metre height but I can also give you a ridge with a 60 degree slope and a 10 metre height so which are you are talking about because when you do not explain clearly, I cannot teach the computer what is a 'ridge'."


Why do some GIS users find GIS difficult to understand?
It has observed during training that many GIS users, both first timers and regular users alike, merely learn how to apply GIS but seriously they do not question the logic behind why each step is required to do a specific task i.e they just take things for granted. This is especially so where a country's education system indoctrinates students not to question their teachers but to accept whatever is taught. How sad. Thus, when such GIS users make a mistake, they find it hard to comprehend why a mistake occurred (must be the GIS software, can't be them) subsequently, it becomes difficult for them to resolve that problem. For example, when corrections in the attribute table need to be made, some users do not see why they need firstly to click the editing icon in the first place because what they intend to do is an editing job. Another issue is that many do not like to explore a GIS software to find out if a specific task can be done with that GIS software at that moment of time (a GIS software is a dynamic tool that keeps on improving as time goes so what maybe not possible now maybe possible in the near future because someone contributed a module in the case of OSS GIS). To jog their mind is a difficult mental task for some because maybe they like to be spoon-fed when they were a toddler, who know?


What is the difference between the GIS and the modules?
  • A GIS is the basic software and can handle basic queries and analysis. It cannot do everything especially the more demanding queries.
  • A module is a supporting tool to that GIS sometimes pre-installed, other times, must be acquired and installed 'on top' of that GIS software. For examples, for 3D vizualization, get a 3D spatial analyst module, for origin-destination analysis, get a network analyst module and non-3D spatial analysis, get a spatial analyst module. Proprietary modules are not come cheap and can cost about half to the equivalent of a basic GIS software. They are also not shareable between proprietary GIS software but the opposite with Open Source GIS i.e. not only are they free but shareable between Open Source GIS e.g. QGIS and Grass GIS modules.

Can QGIS be installed at a server?
No. Well...not the desktop GIS most people are familiar with but yes, a team out there developed a QGIS server. It is still in its youth so I think it will have a hard time competing with the likes of an established enterprise geospatial database such as the PostgreSQL-PostGIS. If GIS is needed at the server level using Open Source, the strategy would be to install PostgreSQL-PostGIS. PostgreSQL is in the league with the proprietary software Oracle, just that it is free. PostgreSQL (mother) is the main database software and handles the non-spatial database while the module PostGIS (son) handles the spatial database. People NOT interested with geospatial database but merely require a non-spatial database would be adequate with PostgreSQL. The GUI-friendly PgAdmin helps administrate the server from a client computer in things like create a new database, create new users and upload files. The GIS user then works from the client computer using QGIS to access the geodatabase in the server. QGIS and PostGIS follow open geospatial standards so work seamlessly, in fact, QGIS even has a default PostGIS icon in its menu bar. Cool! If there is a need to get all that geodatabase "internetable", then invest in a Open Source web server software such as GeoServer and  MapServer to export that data to the Internet. Kevthekelt said that "When you use QGIS in conjunction with Geoserver/Postgis/openlayers/python/Geowebcache you have a setup that not only matches the ESRI desktop and SDE but actually OUT PERFORMS it!" How's that for you?


Why is GIS getting popular?
This is a material world craving for power and wealth thus information is so powerful, it can replace goods and services to deliver that. Traditional information were non-spatial kept in spreadsheet like balance sheet and relational database. People now realize beyond non-spatial information, spatial information can give more power and wealth. The IT tool for that is GIS so many public and private agencies now develop their respective geospatial information. Even McDonald uses GIS to locate the fastest route to deliver their products to their clients in America. That is why GIS is getting popular.

There are quite a few for the iPhone but QGIS already has one for the Android. I'm keeping tab on this subject here. We need to give it some time to mature but hey...that's QGIS for you. Very up-to-date mind you! I would expect some exciting news from Ubuntu land since Ubuntu over Android operating system will become operative late 2012 and this could have a dramatic impact on supporting application such as QGIS. We have to consult QGIS developers for that so join the QGIS Users Mailing List if you are that interested because from my experience, feedbacks are less than 24 hours which is often faster than a formal contractual support from consultants (my experience again).Is there GIS Mobile?

How can GIS benefit me?
Money, money, money. What is money to you? If GIS analysis of your business is properly executed, it will help guide among others, how to save money by optimally distribute branches base on their catchment area, reduce maintenance cost, expedite transport of goods within the shortest time possible and identify potential areas of opportunity. GIS exploits the spatial element not present in non-spatial database thus widening the scope of information to the agency. If cost is a serious factor in your agency, the use of open source GIS outguns proprietary GIS hands down.


What is a GIS software?
A software that can transform geographical features into a digital world. There are 2 versions of GIS: Close Source and Open Source. Both are built on different philosophy which in the end, mould users to certain personalities, for example, Open Source users tend to share knowledge because they depend on it for their survival thus tend to be friendlier and always willing to share, even among strangers where the most probably reason they do not is only because they do not know the answer. This is most obvious when customized modules are built on Open Source and sharing it does not impose any cost to the new user whereas the same cannot be said of Close Source modules.

The most popular Close Source GIS is arguably ESRI's ArcGIS as they have a 20 year start. The file extension of this proprietary software known as ".shp" or shape file has become the unofficial industrial standard. MapInfo, another  proprietary GIS, over 7 years old, is popular for its mapping capabilities and a cheaper alternative to ArcGIS. It uses the ".tab" file extension and tends to venture into niche areas left out by ESRI products and arguably, is a better desktop mapping tool as it promotes itself as one. Open Source GIS has been around sometime e.g. GRASS GIS was developed by the US Military in the 60s and recently others have mushroomed (QGIS's first stable release only came in 2009). Open Source alternatives is getting popular. Firstly, Open Source GIS is freed. Secondly, developers have full access of the source code. Quantum GIS (QGIS) is probably the most popular because it is multi-platform, multi-language, user-friendly and can borrow many modules from GRASS GIS. However, not all vector models are the same: the GRASS GIS vector model is a slightly different approach as against the older ESRI Simple Feature Vector model. To illustrate, all types of vector features can be placed on a single layer under the GRASS GIS vector model but have to be placed in separate layers under the ESRI vector model. Secondly, a boundary between 2 polygons in a GRASS vector model is shared between them thus the area for the polygons gets located at the polygon's centroid whereas but a boundary between 2 polygons in the ESRI vector model is really an overlap of 2 boundaries thus the area of the polygon is located in the actual space of the polygon. The GRASS GIS vector model arguably has a better grasp of the real world situation though the ESRI vector model is more popular because of the legacy it entrenches in many agencies that use it. Thus, if data-sharing between agencies is desired yet among them use different vector models, then data-sharing may prove to be difficult.

Which should I choose: Open Source or Close Source GIS?
It is not a question of which is better but more importantly which of the two can meet your specific  objectives, remember different people have different requirements so you are not obliged to follow anyone and neither is anyone obliged to follow your choice. But if you ask me, it is wiser at first to give OSS GIS a try, in fact, I dare you to try it! There's no harm done and if you have tried it and don't like it, just uninstall it and "throw it in the dustbin". It can't be that bad can it? After all it's free! Don't let prejudice muddle your thinking. Now if OSS GIS still cannot fulfil your objective then obviously you have no alternative but to choose proprietary GIS, like it or not, you have to folk out money. At least you have made a comparative judgement and that makes you an expert on this specific matter but of course if you analyse objectively and do not let prejudice cloud your mind which can be easier said than done.

Is GIS cheap?
What is "cheap"? This is a relative question and there are licence (proprietary) and free versions. GIS, like computer-aided-designed (CAD) software, is  a specialized software so proprietary GIS is not cheap as a software, in fact, they are really expensive because it can cost more than the price of 2 off-the-shelf computers in Malaysia. Gulp! OK, the standard GIS software is adequate for viewing and Standard Query language (SQL) query and when the user furthers into GIS analysis, some pre-installed modules or plugins maybe adequate but when the user advances to heavy-duty GIS analysis, the appropriate module or plugin may not be available in proprietary GIS and has to be acquired as a supporting software. Proprietary modules are not cheap either where the popular type such as spatial analysis can cost the same price as the standard GIS software. In open source GIS, modules are free and shareable where in the case of 400+ GRASS GIS modules to QGIS.


Can GIS software accurately represent the real world? 
I don't think current GIS technology can. You'll have to wait till until technology reaches the stage of Avatar. (Just kidding). One translates the real world situation into GIS either using a raster or vector format and in this respect different users may require different needs to answer their questions and sometimes raster format answers their queries better while in other cases vector format is the better alternative. I believe many people relate the real world situation better with the vector model. Having said that which vector model does one use? The popular ESRI vector model uses a simple approach separating point, line and polygon features which represent specific geographical features in different layers. On the other hand, the GRASS version allows those different vector features to be on the same layer. This sounds closer to the real world situation don't you think so? The problem is that many agencies using GIS have a legacy of using the Simple Feature Vector Model so migrating to the GRASS vector model may be hassle. Furthermore if data-sharing between different agencies that uses different GIS format is taken into consideration and may one day be a reality, it is better to choice the alternative entwined with the older legacy just like it best to keep UNIX rules when establishing GIS metadata.

Current GIS software can show 1:1 relation well such as 1 polygon represents 1 landuse activity e.g. residential. The problem appears in 1:many situation where 1 polygon represents 2 or more activities e.g. a polygon containing an apartment with with commercial activities on the first 3 floors and residential dwellings from the forth floor upwards. Times like these, that polygon needs to be hyperlinked to a relational database.

Another problem is from plan view the GIS shows a underground utility pipe to be 1km long but in actual fact that pipe in side elevation could be sloping and be 1.3km long. So if that 0.3km difference was not taken into account who is going to pay for the additional cost of implementing that 0.3km or was it ever taken into account in the first place? The same can also be exemplified when a road down a hill slope from plan view is 10km long but from side elevation is actually 9.5km. The GIS software will die stating that the actual distance of the road is 10km but who is going to tell the GIS software it made a serious mistake especially when it concerns the annual maintenance cost of that road.

What is GIS used for?
Other than the obvious (view GIS data), it is used for mapping and GIS analysis purposes. How GIS is actually used for varies among the users but the common denominator is the spatial element.

Can QGIS be Web GIS-ed?
Firstly, get a clear understanding of the definition of webGIS as mentioned above under "What is web-GIS?" or else we could be on the same topic but on different frequency. Secondly, to answer that question...No. Here again... not directly. Many people think that Web-GIS is a direct mirror of GIS capabilities of a desktop GIS. Actually, GIS data created by the help of a GIS software gets stored in a geodatabase server then the selected data desirous to be uploaded to a GIS web server. Applications specially designed to extract certain information from that data is usually kept in a web application server and used together with the web server. Most GIS consultants I know would not recommend online GIS editing because such work tend to be 'heavy' on the seerver since GIS data tend to be large files. The problem is that some people refuse to accept this and when things go sour they will say they were ill advised.

What is the connection between data collection, mapping and the GIS analysis part?
To spatially analyse geographical issues, a map is needed. However, the type of data to be digitized in the map depends specifically on what one wants to know. Many people know what features make a map but are bewildered when asked specifically what type of data is needed to answer what they want to know, i.e. they don't really know what they want and and they expect the GIS officer to do their thinking for them which is impossible to do since values are personal (i.e. the value of my $1 can be different from yours) but in reality this is usually the case.

How did data get stored in GIS?
GIS deals with spatial data and can be stored in a non-relational or relational database. Gulp! In the first place, what is a non-relational database? Well, a non-relational database, often called a "flat file" is where records that do not have any relationship between one another are kept. I do not know why a flat file is so called 'flat' but I guess this was meant to differentiate from records stored in inter-related tiers like an organisation chart hence the 'taller file' :-D. An example of a simple database is a flat file and has data stored in a table where records are kept in rows with the common features (fields) stacked by columns where a datum is a single record. In GIS, this represents the non-spatial attributes of spatial features leaving the spatial data to be kept in the form of plans. Things are fine for a flat file when it concerns 1:1scenarios but in life, we are sometimes faced with 1: many scenarios like a plot of land with 2 landuse activities e.g. an apartment with commercial activities below and residential dwelling above. A flat file cannot handle such situations and this requires a more complex database namely a relational database. Since current GIS are based on  the 1:1 concept, in situations which involve 1:many scenarios, the user tends to compromise and decides which is the major element. If we refer to that apartment example and more floor space was taken up by residential dwellings, then that plot of land is recognized as 'residential land'. Of course, it does not paint a true picture of the real situation. To accommodate that commercial element, the spatial layer actually needs to be dynamically link to a relational database. Now whether the GIS user does that is another question and I will not be surprised if it is not either because people do not know how to or just want to keep it simple but the point I want to stress here is that there are still limitations in current GIS software in relation to real life situation. Relational database are usually kept in management system software so we get the term 'RDBMS' which is short for relational database management system and to further make things more complex, there is the usual RDBMS and the spatial RDBMS of which the latter also caters to syncronize between non-spatial and spatial database. An example of an enterprise RDBMS is Oracle while its spatial RDBMS is Oracle Spatial whereas the Open Source equivalent is PostgrSQL and PostGIS. Mmm...this is getting to be heavy stuff!

What plugins are available in QGIS?
This is a difficult question to answer because the issue is dynamic and changes over time when plugins get contributed by some kind soul out there. Today, one figure, tomorrow another so the user has to updates the plugins under Plugins/Fetch Python Plugin.../repositories/3rd party repositories. When I first started with QGIS, it was 129. It is now more than 180 plugins, tomorrow who knows?

What is GIS basics?
In the ESRI Simple Feature Vector Model, just know there are only 3 features that can make up a map: point, line and polygon and they translate a real world feature. Each feature requires a separate digital paper (commonly called layer) and when overlaid, they can offer a better composition of the real world. Secondly, each layer must have the correct map projection or else layers cannot be properly aligned when overlaid for viewing and analysis purpose. This model differs from the GRASS GIS vector model explained earlier above under "Can GIS software accurately represent the real world?" 


Why should I migrate over to a geospatial database?
  • Concurrent use of the same layer possible by many users;
  • No layer gets locked when someone runs off to a meeting or goes missing in action;
  • Database gets automatically catalogued or indexed;
  • Panning and zooming in or out is faster;
  • Query and analysis is faster;
  • Processing does not hang;
  • Security is better as data is centralized;
  • Data-sharing is possible due to centralized inventory;
  • It is easier to manage, control and coordinate a database of voluminous size and diversity;
  • Simplify data management for conversion and uploading to webGIS;
  • Less headache for the GIS technical personnel.

What is 'map projections'?
It is the scale of a map being projected onto the GIS map in relation to what it is really on the ground.  The closer to the ground you come, the smaller the scale becomes.  Take 2 similar torchlight and mark an X with a marker on the face of its glass. Now shine the light on a wall where one torchlight is close to the wall than the other. Note that both Xs on the wall are of different size. This is because they were projected at different distance to the wall. Similarly map projections implies the scale was projected for you to see where the more you zoom in for a closer look means 1 inch on the map gets closer to the real 1 inch on the ground. Mind you, if you want to get more confuse or be more clear read on to the next...


How are things mapped out in GIS?

A geographical feature is mapped out from their  longitude (long.) and latitude (lat.) values in a Geographic Coordinate System but when it is place in a Projected Coordinate system, that feature is projected in a GIS layer or in layman's term: onto a digital paper.  So you will know by now there are 2 types of coordinate system. The long. reference becomes the x value and the lat. reference becomes the y value like on a graph chart we learnt at school and is expressed in degree where the coordinate can either be positive or negative. Long. values are positive in the eastern hemisphere but negative in the western. Similarly, lat. values are positive north of the equator but negative in the south. The fundamental difference between the two coordinate system is that the Projected Coordinate System has constant length, angles and areas between 2 dimension whereas this is not the case in a Geographic Coordinate System as you should know earth is really not a sphere and is a bit fat around the belly where land forms can be high or low here and there. A datum or point of reference for measuring geographic coordinates can at times use a Geographic Coordinate System or a Projected Coordinate System. For example, Kertau is a datum used for RSO topographic work in W.Malaysia but different states use different datum in their Cassini maps i.e. there is one in Perak and another in Johor. With the projection issue cleared, their non-spatial data are linked to an attribute table. GIS layers can be in vector or raster format. The problem as I mentioned earlier is how do you map out things in a 1:many scenario ? It is easy for a 1:1 scenario because a point, line or polygon is associated to their attributes. If you try to cheat by taking the major landuse or land activity for that feature especially where it concerns  a polygon, you have left out vital information in your data collection and hence your analysis will not represent the actual situation so who are you trying to fool but yourself?

How to work under 3 map projections in GIS?
Let's get this straight: this is not an issue concerning a GIS software be it open or close source. This is fundamentally the problem of uniformity and integration so don't go barking up the wrong tree (or wrong agency). Alternative 1, you decide which of the 3 map projections will be the main projection then re-project the other 2 projections to follow that main. If there are errors (I wouldn't be surprise if there isn't some but the degree of which I do not know) there is nothing much you can do. Alternative 2 is choose a projection different from all 3 but this only increases the probability of error so I'd choice the former though which of the 3 should be the main projection I would seek advice from the map custodian and here I'd expect it to be a state casssini map. You can't ask for everything when you depend on others for data but beggers can't be choosers so don't complain but make the best of what there is.


What is 'vector' format?
A fancy word to impress upon people  so that you sound like a neutron scientist while at the same time, represents a format for presenting data base on lines such as cadastral maps. Point, line and polygon are the 3 basic vector features that represent real world geographical features each getting more complex or powerful than the other

What is 'raster' format?
Another fancy term to impress upon people so that you sound like a space scientist while at the same time represents a format for presenting data base on a grid of cell or pixels such as a photograph, tv or satellite image. The shape of  the pixel is usually square so as not to distort the image it represents. Images taken by satellite use the multi-band of light which are translated to represent a geographical feature. Natural features like different types of crops show up well using raster whereas the different types of landuse activities in built-up areas do not show up well using raster.


What is 'topology'?
Ah..another fancy word people is in the GIS world use but dare not explain what it actually means. Simply put the way to understand how real world features are transformed into a digital world so that they can be representative of the real thing. This calls you to do some creative thinking I mean a line logically cannot be a road but in GIS it can be meant to REPRESENT a road, river or a railway line if need be. To exemplify, for a line to represent a road often means many segments of lines that are joined together correctly meaning their ends must meet correctly because when they do not meet or overshoot, that road gets broken up into 2 roads. Also, when 2 lines that represents roads cross each other but no node was created where they intersect, that would imply there is a flyover there but which line flyovers  which, current GIS cannot tell. If the data of the roads were input then selecting a line will help tell data of that particular road. When a node was created at that crossing, it implies a cross-junction but whether there is a flyover there, who knows? The thing is that, it becomes now possible to measure the distance of a route as it criss-cross lines. So in short, the topology of a vector feature becomes more complex when more is required to make a specific vector feature where the simplest feature is a point and becomes more complex when it becomes a line and most complex when it becomes a polygon meaning the probability of topology error increases when that feature becomes more complex...tough luck!

What is 'topology error'?
Error in the creation of a digital feature. Since a point is merely a dot, I don't see how you can make topology error but  as the digital error becomes more complex the degree of error increases i.e. all things being equal, a line feature like a railway line comprising of many segments of line can have less segments than a many sided polygon which is the most complex of the three to create.  When drawing line features like a railway line, a crossing of lines with with without node means something different so get that wrong and our analysis that depended on that line will be flawed. Line in terms of road can be made one way or two way direction and if this was not corrected input, walla! you get another mistake. The meeting of polygons and lines are common areas of human error. Anyway I guess it refers to errors resulting from inaccurate geometry due to someone not paying close attention, plain lazy or in need of  better spectacles. If a line did not meet at a supposed-to-be junction of another line, it is called an under-shoot and where it exceed, it is called an over-shoot. Obviously, the former needs to be extended whereas the latter needs to be snipped off or shorten. Where two sides of a polygon were supposed to meet but did not, there will be a gap or what some term as silvers because they look like that 'silvery' insect found in old books. Auto- cleaning nice as it sounds cannot solve this problem accurately because the rules applied by the GIS software may not meet your requirement as what is happening on the ground. Splitting the gap 50:50 sounds fair but 'fair' does not meant it is accurate. Even if you go to the ground you will not see the error unless it is a physical boundary, you need a theodolite to solve this! Then are situations of overlap when polygons overlap each other when they are not supposed to but the question of which polygon is right or wrong  is going to be a big headache to rectify.

What is the most common mistakes GIS learners do?
They have a habit of not using GIS software following procedures that need to be done sequentially step-by-step but tend to jump steps. For example,when configuring a vector layer, the learner after selecting which feature is desired then skips configuring the coordinate reference system (CRS) or after naming a new attribute then skip configuring its type and goes straight to configuring the width. Another common mistake especially among proprietary GIS users is when wanting to activate a certain vector layer, say, a .tab file, starts to complain QGIS does not show the layer they wanted only because they did not pull down the file extension type to swap between .shp and .tab. Thirdly, learners complain QGIS is no good simply because they did not switch over to edite mode when what they want to do is an edite task meaning they do not know what they are doing.

How to get quality data?
There are 3 types of data error: topology error, data structure and attribute error and they are all human error. Reduce human error and data quality improves, in other words concentrate and don't fall asleep on the job. What do you expect me to say?


What is data structure error?
Errors resulting from inaccurate data structure when compared against a reference manual. Of course, if you do not have a reference standard to lean on, technically, you won't have any data structure error. ;-D


What is attribute error?
Errors resulting from inaccurate data entry by human error or against a configuration requirement e.g. maximum number of characters allowed

How to reduce data error? 
Wake up! Drink more coffee! Follow closely a reference manual which GIS users follow to reduce data structure and attribute error. Simply don't be lazy.

How to do GIS analysis?
Be clear of what you actually want. Sadly, simple as it may appear, many people do not know this. Remember a problem requires the right set of data to help solve it only then collect it. Many people collect data but are not sure if they collect the right set of data for the problem. Sort out how all the right set of data inter-relate meaning a GIS technique must be ascertained and as usual different type of problem require different approach. Some problem merely require spatial data, others a mix of spatial and non-spatial data. Your human values must input into a matrix especially when it concerns 2 sets of spatial elements to be considered. These inputs teach the GIS how to translate that graphically....Oh my, this is going to be a difficult question to answer.


Which is the worst area in to tread in GIS?
Analysis. Why? Because of 3 factors:

  • The probability of inaccurate results is directly proportional to the amount of error in the data. The issue is how certain are you that your data is in tip-top condition?
  • The probability of inconclusive results is directly proportional to the amount of crucial data not collected in the first place. Many people want GIS information but the issue is that they did not do the appropriate data collection in the first place probably because they do not know what prerequisite data are needed  to analyse a particular GIS issue.
  • Many people do not know how to come up and justify parameters and more so, how to justify value judgement for the various parameters. For example, people may come up with 5 grades for exam results but the issue is that they do not know how to justify why a passing mark can be 36 and not 41 or 51. This weakens the analysis.
Last but not least


How to explain to a GIS illiterate yet get the message across to this Dopey?
When you look through the eyes of that person, do not think he or she understands eventhough that person can be your top management. Think now of that person as your typical sweeper. Yes! I mean it...downgrade that fellow even if he or she is No.1 because it will also give you the spirit to energize you to tell that person a thing or two where you excel and he or she doesn't. The moment you see or better still, believe, that person as a sweeper, your mental approach automatically changes and you become more analogic when trying to find examples you found hard to find before. This makes you more creative in finding ideas to express yourself. The other point I wanted to say is that if you can make a person who literally does not have the tiniest grasp of GIS which is definitely not in his field yet later he says he now has a clear understanding of GIS eventhough it maybe basic GIS, then you must give yourself a tap on the shoulder! And, of course ... never reveal this secret or else you are asking for it :-D

31 comments:

  1. 200% Aggree with u SIR. Thumb UP! Our difficulty RISEN, when we are facing with human factor problem. BIG Problem can be solved when people UNDERSTAND each other. If not, the BOSS is always RIGHT? We are in TROUBLE!

    ReplyDelete
  2. En. Abbas, currently I AM SO DISAPPOINTED WITH MY ORGANISATION and my TOP MANAGEMENT. I'm thinking to quit and and let them handle by themselves. From the bottom of my heart, i really LOVE GIS, But people around me, the environment and FUTURE and not supporting me. I am really sad when think about this and my future. :(

    ReplyDelete
    Replies
    1. Don't be. Life is just a game. Somedays we win, others days we lose. Just like a lion, the king of jungle, yet who just lost territorial rights against another, he will always lick his wounds to fight another day. We are no different and we must follow that example. The challenge is one has to be more creative and innovative in finding ways to convince and win their heart no more different than a salesman trying to secure a deal. Top management just need more convincing. For example, I wanted to promote Ubuntu operating system in my department and although a live demo got them talking it was not until I convinced a staff whom a director knows that he is IT savvy of the benefits of Ubuntu and after that guy did the convincing for me that director now uses Ubuntu on a daily basis and brags about Ubuntu.

      Delete
    2. Anon, En abbas..

      I'm also have the same problems. Quite sad with them but i have to make them understand bout GIS even though there's many negative thinking about our Unit. Really sad cause nobody support us, sometimes am also thinking to quit but this GIS is one of my passion..so we need to convince them, win their heart for accept us. Yup, it's really really really hard but i will never give up...hopefully...

      Delete
    3. Never mind, your passion will pull you through as you soon realize you just came back from the future therefore cannot blame them when every time you show in your hand, you hold a cube, they can only see a square. That is the price of being a pioneer.

      Delete
  3. En.Abbas, i know what u mean & i hope i can be like u 1 day.BUT, if J54 u present A TALK/PAPER to top management (JUSA etc), they will say u are SOMEBODY..BUT If Grade 48/44/41 it is nothing..that is from my experience, IT is very challenging to convince TOP MANAGEMENT WHICH has so much EGO to admit our TECHNICAL OPINION!

    ReplyDelete
    Replies
    1. Now, now, forget about rank and file. Concentrate on what is the selling point of your case e.g. saving money. Then address every criticism but constructively until they run out of gas. Take it like a game of chess and don't let your opponent even "check" you. Apply "opposites" what you learnt at school, for example the opposite of ego is flattery but the game is when to and when not to apply. Strategize... confrontation between 2 parties is best by a third party. My junior officers know I had a difficult time as a J48 with my superiors and it was not until I won the MAMPU Open Source Award that they now listen when I have something to say. To gain their respect, in meetings you must also be able to pose important/crucial questions that even they did not see and do not know how to answer which is what I do, of course I do not get many true friends but that is the price. Is it worth the risk? You ask me.Decisions, decisions.

      Delete
  4. WOW, you really ENLIGHTEN me now. I hope i will try wo WIN this game. My new SLOGAN is "KNOWLEDGE & SMART IS THE POWER". With knowledge, i can win all kind of DIFFICULTIES.
    Thank you again SIR.I love u.

    ReplyDelete
    Replies
    1. And when life is painfully difficult, take a break and sing Nat King Cole's song "Smile". Take care.

      Delete
  5. what mean by attribute information? can u give me some example?

    ReplyDelete
    Replies
    1. The technical specs of a particular attribute,e.g. the max no. of character for the name of a feature is 15characters. This has direct implication for certain GIS software.

      Delete
  6. nak tanya. Apa kepentingan analisis ruangan dalam GIS? Boleh bagi contoh sekali?

    ReplyDelete