'Learn from Map' application (Learn Geography)

[Varun Sashidharan]

Greetings!

I am Varun Sashidharan, a 2^nd year Computer Science Undergraduate student at PES Institute of Technology, Bangalore. While I was going through the list of projects, the project titled “Learn from the Map”, particularly interested me as I’m currently working on another application dealing extensively with maps and Geo-location.  Currently, I’m using the Google Maps api V2 and building my app for the Android platform.

I am comfortable with Android having worked on 3-4 apps in the past year, and attended several “Hackathons”. I have recently started learning how to develop apps for Windows 8 using Micrsoft Visual Studio, though Android remains  my preference.

I am familiar with the following programming Languages -C,java, Basics of python, Assembly, Verilog and R. I have also completed courses on Data Structures and databases as part of my Undergraduate course.  

As regards to the project, I propose we build a server which stores the details of all users’ previous log-ins, and enable each user to play with his friends across the internet. This would lend some competition into the game, and from personal experience, I’m positive that children would grasp knowledge better and faster when they want to beat their friends at it!  A good example of this, is the popular game “Quiz Cross”.

Also, we could develop the app such that it can target different age groups. For example, for a 5th standard child learning about Rainforests, we can ask the user to point out various regions on the map likely to have rain forests and then ask related questions. However, for more mature students of geography learning about specific tribes, location of minerals, weather currents etc( classes 9th-10^th), the Questions can be more detailed. I'd like to know what age groups are being targeted, as this would help me think of some ideas specific to those age groups.

As regards to the problem about automating questions, I believe a database would be essential, so that we map the questions based on the key values specified. We could also, specify the difficulty levels for a range of questions, mapped with the region and maturity of the user. Also, if you are looking to globalize this app, based on the location of the user, questions can be automated. For example, questions on USA’s history, will have a lower index if the user is in India.

I am a quick learner and would be happy to contribute to this great initiative. More importantly, this would provide me with a great platform to channelize my self-taught android skills into something useful. I believe I can dedicate my time to completing this, and will work hard to make it a success.

Regards,

Varun Sashidharan

Email id- svar...@gmail.com

2^nd year, PES Institute of Technology

Bangalore.

[Rahul Ahuja]

Hi Varun

Thanks for introducing yourself and your skillset.

As regards to the project, I propose we build a server which stores the details of all users’ previous log-ins, and enable each user to play with his friends across the internet.

For the scope of GSoC, we are not keen on making this a multiplayer game. To start with, we are looking for a single player game experience without any server dependencies.

 

Also, we could develop the app such that it can target different age groups. For example, for a 5th standard child learning about Rainforests, we can ask the user to point out various regions on the map likely to have rain forests and then ask related questions. However, for more mature students of geography learning about specific tribes, location of minerals, weather currents etc( classes 9th-10^th), the Questions can be more detailed. I'd like to know what age groups are being targeted, as this would help me think of some ideas specific to those age groups.

The target group is students in primary school who study geography. I think you can have a difficultly level defined in your game which handles the mature age groups as well.

As regards to the problem about automating questions, I believe a database would be essential, so that we map the questions based on the key values specified. We could also, specify the difficulty levels for a range of questions, mapped with the region and maturity of the user. Also, if you are looking to globalize this app, based on the location of the user, questions can be automated. For example, questions on USA’s history, will have a lower index if the user is in India.

Yes, you need semantics and meta-data atleast for geographical locations. You may or may not use a database for it. We would recommend reading this meta-data off an XML file bundled with the application.

 

I am a quick learner and would be happy to contribute to this great initiative. More importantly, this would provide me with a great platform to channelize my self-taught android skills into something useful. I believe I can dedicate my time to completing this, and will work hard to make it a success.

Great, try to crystallise your idea more and work on a proposal which you may then have to submit to Google.

[Varun Sashidharan]

Hi Rahul,

Thanks for your reply.

In addition to the quiz i thought of a few we ways we can improve the functionality of the app. I'll cover these below.

The target group is students in primary school who study geography. I think you can have a difficultly level defined in your game which handles the mature age groups as well.

Since you said that the target group, is students in primary school, I thought we can develop the app such that we can allow the users to interact with the map, by drawing shapes and lines. These can be compared with the correct regions/shapes, thereby enhancing the learning experience. Directional skills and simple understanding of maps can be taught to primary students by showing them the routes by which they traverse.  

Yes, you need semantics and meta-data atleast for geographical locations. You may or may not use a database for it. We would recommend reading this meta-data off an XML file bundled with the application.

I agree that storing information as meta-data may be a simpler solution and occupy less space. But, to make the app powerful enough to handle a wide range of queries and constraints, which match attributes from different tables a database could be used. I believe that in addition to asking standard questions as part of the game, the app can be used as a comprehensive tool to enrich learning. Visual representations on the map based on a cross referencing different attributes may be easier and faster to comprehend. 

For example, to select: a rain forest, with an indigenous tribe, and a region of agriculture. A representation matching all the attributes should show on the map.

 Matching keys can be used to generate appropriate questions as well, that can be generated based on the user's difficulty level and region of interest. This is would be extremely useful for more mature users.

Also, in addition to google maps,  we can load a set of standard maps that are usually used in the school level. Students may directly practise marking locations, capitals, etc. This would reduce the dependence on paper, as well as provide them with a ready set of maps that they would need to know. So, once the user marks/draws the locations,we can then display a "Solution Layer", which shows the students the correct locations as well as gives them an idea of how accurate they were. 

I'd like to know, how this tool would be used. Will the users work on a tablet/phone of their own or ones provided at the institution? If the gadget belongs to the user, we can incorporate basic AI to test the user based on his previous results.

I've attached a few screenshots of a basic tracking app that i built using GoogleMaps api V2. The app has all the basic map layers, including the user's location, zooming in-out, tilting, compass, etc.

 

Looking forward to your reply,

Varun Sashidharan

[Pankaj Nathani]

Hi Varun,

 

you said that the target group, is students in primary school, I thought we can develop the app such that we can allow the users to interact with the map, by drawing shapes and lines. These can be compared with the correct regions/shapes, thereby enhancing the learning experience. Directional skills and simple understanding of maps can be taught to primary students by showing them the routes by which they traverse.  

I like the idea of allowing to draw lines and shapes on the map. But remember, the goal of the application is to teach general geography and not map directions, etc.

 

I agree that storing information as meta-data may be a simpler solution and occupy less space. But, to make the app powerful enough to handle a wide range of queries and constraints, which match attributes from different tables a database could be used.

You are free to use as many datasets as you want. Just that there's no server integration for now, so these datasets have to be local.

Also, in addition to google maps,  we can load a set of standard maps that are usually used in the school level. Students may directly practise marking locations, capitals, etc. This would reduce the dependence on paper, as well as provide them with a ready set of maps that they would need to know. So, once the user marks/draws the locations,we can then display a "Solution Layer", which shows the students the correct locations as well as gives them an idea of how accurate they were. 

I like this as well. The question is where would you get these standard maps from? May be there's a way to trim down google maps into bare minimum school maps? Not that I know of.

I'd like to know, how this tool would be used. Will the users work on a tablet/phone of their own or ones provided at the institution? If the gadget belongs to the user, we can incorporate basic AI to test the user based on his previous results.

The users would be primary school students studying geography. Mostly they would be using it on a fairly large screen (4 inch+) phone or tablet. The gadget may or may not belong to the user, most likely it would belong to teachers or their parents.

Regards,

Pankaj.

 

[Varun Sashidharan]

Hi Pankaj,

Thanks for the reply!

I like the idea of allowing to draw lines and shapes on the map. But remember, the goal of the application is to teach general geography and not map directions, etc.

I understand. The concept of drawing can be applied to the quiz questions as well. A set of questions involving pointing out different regions such as a mountain range or forest, can be answered by simple drawings. I suggested this because Ii think it will appeal to younger kids as it increases user interaction.

 

You are free to use as many datasets as you want. Just that there's no server integration for now, so these datasets have to be local.

 

Yes, without server integration,storing a large number of datasets may pose problems of storage as well as cause unnecessary cluttering. This however can be incorporated if ever a more advanced version of the app is made , to cater to students from different parts and belonging to different age groups.

I like this as well. The question is where would you get these standard maps from? May be there's a way to trim down google maps into bare minimum school maps? Not that I know of.

Standard maps, need not be google maps, or satellite imagery. They may be simple uploaded images. From personal experience, I can recall that geographical maps used in the primary-middle school levels are available in most stationary shops and are just simple outlines of regions with rivers,natural boundaries, etc.  A scanned and appropriately photo shopped version of these maps can be used. Anyway, i believe that Google maps are the way to go, but for the exercise of drawing and practicing on the map , alternative maps may be displayed since Google Maps provide an abundance of data that is irrelevant to a young student, which might be cause for distraction.

These ideas are merely suggestions to improve the capabilities of the app over and above the existing functions, if and when the app is developed on a more global scale. 

The quiz can be a set of pop-up questions that are asked by a cartoon character resembling some animal/inhabitant belonging to that region. I think the success of the app, depends largely on how well it appeals to the target audience.

Looking forward to your reply,

Regards,

Varun Sashidharan

 

[Pankaj Nathani]

Hi Varun,

  

I suggested this because Ii think it will appeal to younger kids as it increases user interaction.

Sure, you are feel to use any map interactions but not to teach map directions is what I meant.

 

Yes, without server integration,storing a large number of datasets may pose problems of storage as well as cause unnecessary cluttering.

You may not want to store complete questions.. you may just want to store some meta data about geographical regions and let the app form questions for these regions.

 

 A scanned and appropriately photo shopped version of these maps can be used.

This is easy, the difficult part is mapping geo-coordinates (and consequently locations and boundaries) to this scanned map. 

The quiz can be a set of pop-up questions that are asked by a cartoon character resembling some animal/inhabitant belonging to that region. I think the success of the app, depends largely on how well it appeals to the target audience.

Yes, the more interactive the experience, the better.

Regards,

Pankaj. 

 

[Varun Sashidharan]

 Hi Pankaj,

 Just wanted to ask a few basic questions.

 

You may not want to store complete questions.. you may just want to store some meta data about geographical regions and let the app form questions for these regions.

What about using SQLlite? Would you recommend it, as a suitable alternative to store our data to generate questions? Or meta-data in XML files the way to go?

 A scanned and appropriately photo shopped version of these maps can be used.

This is easy, the difficult part is mapping geo-coordinates (and consequently locations and boundaries) to this scanned map. 

There is a solution to this, however it involves considerable image processing: 

We can first map the geo-coordinates onto the image by applying suitable projections(such as miller projections) and then convert the projection based latitudes and longitudes to actual pixel positions. To test the drawing made by the user with our version of the solution, use a fingerprint based algorithm, where the fingerprint value of the image is tested against that of the solution an image, after the tester image has been re sized and normalized.

This I think deviates from the goal of the project. Or do you reckon it can be included as one of the project modules?

Regards,

Varun