We are now LIVE with v1.0

Although there are still some objectives not met yet, I am naming this release v1.0 as the database would not be going live. No more test data and no more clearing the data. What you plot would be what you see.

This time I shall host at Dropbox as well as at Github.

We are now LIVE!

APK File: https://www.dropbox.com/s/s68ip7clr9fjgso/Fruit3SG.apk

User Manual:  https://www.dropbox.com/s/v5q08aa3h369kjx/Crowd%20Soursed%20Data%20Collection%20Manual.pdf

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

Source Code: https://www.dropbox.com/s/n0vqb6caa6vm10q/Fruit%20Tree%20SG%20Source%20Code.zip

GitHub: https://github.com/Ashwani001/Crowd-Sourced-Data-Collection-


Get the data here:
Data in Webpage: https://docs.google.com/spreadsheet/pub?key=0ApvqX5TtdVUIdGdDU3NWbms4Tm1PT3VVUk9Ya0hYUkE&single=true&gid=0&output=html

Data in CSV Format: https://docs.google.com/spreadsheet/pub?key=0ApvqX5TtdVUIdGdDU3NWbms4Tm1PT3VVUk9Ya0hYUkE&single=true&gid=0&output=csv

Data in Plain Text: https://docs.google.com/spreadsheet/pub?key=0ApvqX5TtdVUIdGdDU3NWbms4Tm1PT3VVUk9Ya0hYUkE&single=true&gid=0&output=txt

Data in PDF: https://docs.google.com/spreadsheet/pub?key=0ApvqX5TtdVUIdGdDU3NWbms4Tm1PT3VVUk9Ya0hYUkE&single=true&gid=0&output=pdf

Data in ATOM: https://spreadsheets.google.com/feeds/list/0ApvqX5TtdVUIdGdDU3NWbms4Tm1PT3VVUk9Ya0hYUkE/od6/public/basic

Breakthrough with v0.7

Finally I managed to load the markers on start up.
Was so happy to see the fruits on the map!
The jump between version 0.6 and 0.7 was the greatest so far; in terms of lines of code added AND how much I learnt.

I had to modify the way I was uploading data. Spot the difference?

However one problem is that the current markers take awhile to load.

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf
APK File: https://www.dropbox.com/s/tbx1m40mvm8tb9o/Fruit-3SG-v0.7.apk
Source Code: https://www.dropbox.com/s/e38kswp2lj3p7qb/FruitTreeSG-v0.7-SourceCode.zip

PLEASE NOTE: Will no longer be hosting files on dropbox from next version. Follow my project on Github: https://github.com/Ashwani001/Crowd-Sourced-Data-Collection-

The Curse of the 80:20

From Wednesday about 9pm till now, Friday 8.13am, I spent most of my waking time working on the code to import the data fro my spreadsheet to eclipse. The problem was that they recently changed the API, they dropped support for Spreadsheets API and now use Drive API. And ALL the examples out there are for the older API. Even a video done by google employees themselves could not help me. This time I looked for help on StackOverflow again but unfortunately it was not answer.

This could very well be the issue that brings the project to a halt( well without this there would be countless godzilian duplicate entries.

Right now I MIGHT have reached a breakthrough. No promises yet. But I am finally importing the data in a format which I COULD be able to use.

Notice the data starting from TIMESTAMP.

Well for now I need some sleep. Hopefully when I wake up I would be able to think more clearly.

v0.6

Lo and behold!
Google ID captured, address reversed engineered from the Lat&Lng.
Now the main thing left is loading the markers with information from column C&D.
After that will be able to work out the smaller details (including the GPS problem, still have not gotten around to fixing it).
IF time permits might be able to launch a browser version for viewing as well.

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf
APK File: https://www.dropbox.com/s/zco9te53f9bew4s/Fruit3SG-v0.6.apk
Source Code: https://www.dropbox.com/s/5vk2bssiplc5xjr/FruitTreeSG-v0.6-sourcecode.zip

v0.5

Turns out the code I got from GitHub was very well written and very well documented as well.
What I spent about 5 hours trying to do on Saturday I did within 5 mins with this code.
Will still have to learn about AsyncTask another time.

Currently it only records the location and fruit name. Date and time are auto recorded by default.

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

APK: https://www.dropbox.com/s/sa3hlj8lv78ds1g/Fruit3SG-v0.5.apk

Source Code: https://www.dropbox.com/s/bztw95ah3k8a4z7/FruitTreeSG-v0.5-sourcecode.zip

Please note all previous versions have been moved. From now onwards only the latest version will be available. If anyone requires an older version feel free to leave a comment and I will make it available.

Async Headache

Yesterday I tried to upload the submitted data to a spreadsheet. I got the Http Request file off StackOverflow. However five hours later I could not get it to work. The problem was,  from what I understand, that Android does not allow you to do anything that takes too long on the main thread.

They have this thing called AsyncTask which is supposed to help you handle multiple threads. However I found it hard to use. After I wrote the code for it for some reason I was unable to execute it. It kept asking me to pass in a variable although I had not defined it to take in any variables.

I scrapped my code for the AsyncTask and looked through GitHub for a ready made one. I found some code for uploading data to a spreadsheet. It does not appear to look AsyncTask, I hope it was not made for an older version of Android.

I will continue work on it on Wednesday now.

With v0.4 comes an almost completed UI

Quite happy with my work tonight. Achieved what I wanted to. Took longer than expected though.
So the 3 previous buttons that I added now change the icon that the user can plant. I have also added a 4th button which lets user toggle the viewing mood. I spent quite awhile on this as previously I had worked with menu, this time it was image button and they both work sightly differently.
But the good news is that users can toggle between the fruits as many times as they need to; from mango to durian to mango to viewing mood etc . It will work.
Also, while users had to press back button after selecting fruit previously, it now returns by itself.

It might be awhile before the next version is out. Now the basic user interface of the app is settled.

Next area I would have to work on is the flow of data. Where will the data be hosted? How to import&export data between an Android device and a spreadsheet? What sort of information do I need to capture? And another big issue, which I might not be able to add for version 1 of the app, is attaching photos to the markers. 

I think sending data from the app to the spreadsheet will be easy, I should be able to tap into google form. The harder part might be accessing the data to form the map if the spreadsheet is private. I plan to make it private so as to prevent people from deleting things. The data will be available for everyone to view from the app (and hopefully browser) anyway. The spreadsheet also should be downloadable for anyone to use. 

There is a little problem with the loading of map when GPS is on. It does not auto zoom in to location when there is no last known location. Will have to remember to correct that as well.

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

APK: https://www.dropbox.com/s/tc11bydeokd9vu3/Fruit3SG-v0.4.apk

Source Code: https://www.dropbox.com/s/nc42chz2e2iq6s9/FruitTreeSG-v0.4-sourcecode.zip

Please note all previous versions have been moved. From now onwards only the latest version will be available. If anyone requires an older version feel free to leave a comment and I will make it available.

v0.3

I chose 3 fruits to start with.

Created the icons for them.

Created a menu to change the fruits.

However the buttons do not toggle anything yet.

ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

APK: https://www.dropbox.com/s/en3eqazpror0m2f/Fruit3SGv0.3.apk

Source Code: https://www.dropbox.com/s/k0aammfhesyja8f/FruitTreeSg-v0.3-sourcecode.zip

Please note all previous versions have been moved. From now onwards only the latest version will be available. If anyone requires an older version feel free to leave a comment and I will make it available.

StackOverflow, v0.2.2

After a few more hours of searching for the problem I decided to enlist some help on StackOverflow.
I posted my problem there and within minutes it was answered, with one of the replies being the solution.

Link to my question: http://stackoverflow.com/questions/23958873/textview-causing-app-to-crash

And the hero that saved the day:

Source Code: https://www.dropbox.com/s/y5512ugypm3okhe/FruitTreeSG-v0.2.2-sourcecode.zip
APK: https://www.dropbox.com/s/8x0bzz9i7axno90/Fruit3SGv0.2.2.apk
ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

Error Hunting

Well I thought I would add in the Legal Notices.
However I kept getting errors I could not resolve. After very long I realized it was because I had given a layout and an id the same name.

After that my app kept crashing. back when I used to work with C++/C the way I looked for error was by putting 'cout's to trace where the program actually crashed.

I tried the same for this but I realized it was not so easy to print to screen. I am not sure how to access the log file as well.

However I started making portions of the code I suspected the error was at into comments to narrow it down. I spent quite a bit of time changing to comments, fixing, comments, fixing until I finally narrowed it down to a specific function called TextView. I'm going to bed now, will continue work on this tomorrow.

Oh yeah the System.out.println() did not work. Not sure how it is supposed to be used, will look it up tomorrow.

v0.2.1

So far I was testing the code by generating the apk, sending it to my phone using AirDroid and then installing it. Then I realized AirDroid allows you to install the app as well. Just now I stumbled across an even faster way; just connect my phone to the laptop using usb and I can test the app from Eclipse itself. It generates the apk, sends it to my phone and installs it much faster than my previous method.

This minor update now allows user to place icons. Users are still able to pan. Check the ChangeLog to view all changes.

Source Code: https://www.dropbox.com/s/ev7l9mh44s5ztt3/FruitTreeSG-v0.2.1-sourcecode.zip
APK: https://www.dropbox.com/s/2igrkagu0gloc2m/FruitTreesSGv0.2.1.apk
ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

v0.2

In this version I wanted the app to automatically find your location and then zoom in.

However when the map starts it does not do that.

When I click on the locate button, it does manage to find
location and zoom in on it as well.

Those two were with my GPS ON.
Let's see what happened when I start the app with the GPS OFF.
I have not typed in any error messages yet so I expect no changes from the first screenshot.

UPDATE: This is what I see as it uses last known location.
 When I clear cache and reboot the app with GPS off it does not load the map.

Well that is enough coding for one night. Will continue working on it tomorrow. I forgot to mention, I am testing the app on a Nexus 4.

Source Code: https://www.dropbox.com/s/cg68bhsvm86fjm1/FruitTreeSG-v0.2-sourcecode.zip
APK: https://www.dropbox.com/s/isrsw06z5d8xr24/FruitTreesSGv0.2.apk
ChangeLog: https://www.dropbox.com/s/4pw1mf4zu2p36p5/Fruit3Sg_ChangeLog.rtf

v0.1

Turns out the problem was that I did not sign the app. You HAVE to sign the app before you can install it. Not sure what you can do with unsigned apks.

Ladies and Gentlemen, the moment you have been holding your breaths for....I give you, FRUIT TREES SG VERSION 0.1!

For now it only says Hello World and opens the world map. But all journeys begin with a single step :)

Everything is open source so here is the source code as well as the apk:
Source Code: https://www.dropbox.com/s/4zyl4og0mxuir5h/FruitTreeSG-v0.1-sourcecode.zip
Apk:             https://www.dropbox.com/s/abu7r87avf8pk3r/Fruit3SGv0.1.apk

Friends in France?

I was provided two urls that might be useful for the project.

http://www.biofortified.org/2012/02/want-to-identify-plants-with-your-smartphone/
The first one was a project with uses image processing to identify plants. Wanting to stay close to the tune of RERO(Release Early Release Often, version 1 would definitely not contain this.

http://www.plantnet-project.org//page:tools?langue=en

This website has a lot of different programs featured on it. One of them caught my eye as it is about geo-tagging plants on user-defined background maps. The page this application was hosted on was in french but google translated it for me. The software itself is only for windows though. Just sent the link to my partner, hopefully the program is in english. Waiting for him to get back to me.

UPDATE: Program was in french. Have a friend who can read french but he is current in the states.

API Key

Last night I decided to restart the app from scratch. A possible cause of error was the API key. Previously I did not generate my own key as the maps package I downloaded had a key. But I realized I would need my own unique key. Still working on it.

UPDATE: API problem settled.

Inspiration for this project.

Struggling with the app I was reviewing why I chose to do this. Thought I would share my inspiration. The following is a Facebook post by a NTU SPMS professor. A summary is provided at the end of this post.

Cheong Siew Ann

Mass Blooms and Climate Change: A Crowd Sensing Project

Flowers are blooming en mass in Singapore right now. It has been suggested that such a mass bloom is the result of climate change. As a scientist, I believe this hypothesis can be tested. Such changes may not be apparent in a single bloom, but may be discovered if we consider the time pattern of mass blooms over many such episodes. If the change occurred in the past, I know how to find when this happened from the pattern of mass blooms. If the change is imminent, I also know how to detect its early warning signatures.

Seeing that climate change is a topic on your minds, and that many has take notice of the mass bloom, I figure this would be a fantastic citizen science project! Citizen science is not new. The American Ornithologist Society has for many years mobilized hobbyist bird watchers to understand changes to the geographical distributions of American bird species, and thus the impact of human settlements. They have even rediscovered a woodpecker species thought to be extinct this way!

The idea behind this crowd sensing project is that plants are sensors of climate change, and rare mass blooms are sensitive to very specific combinations of climatic signals. We are then sensors of mass blooms, because we appreciate their beauty. I hope through people observing mass blooms, which in turn monitor climatic variations, we can learn something about the underlying climate change.

In this project, I would need your help gathering data on mass blooms. The kind of data I am looking for includes:

(1) newspaper and magazine articles reporting mass blooms (for each article, I will need its scan (or photo), the name of the newspaper/magazine, the date of the newspaper/magazine, and the page the article appeared on);

(2) TV news reports, if anyone has access to archives (for each TV news report, I need a video clip of the report (if permitted by copyrights), the name of the program segment, the date of the program, and the time the report was aired);

(3) photographs of the mass blooms (for each photograph, I will need the digital file, or a scan of the photograph if it is an older hardcopy print. For older prints, I need the date the photograph was taken to appear on the photograph, or some documentary proof that it was taken on the date asserted).

From the data, I would need your help to extract the following information:

(1) the date and duration of a mass bloom;

(2) the geographical locations where the mass blooms were recorded;

(3) the species of plants participating in the mass blooms.

On my part, I will provide regular updates on the progress of the project to all participants. I might also post more specific calls for additional data where I have identified gaps in the overall picture of the mass blooms, whenever there is a need to do so. For now, people can join the project by responding to this post. Please help share this post with friends you think will be interested to contribute. The crowd sensing aspect of this project is novel, and will have to be reported if we ever write this study up for publication. This means that I will need to know who invited you to this project. This information will eventually be published in the form of a graph, where nodes are participants, and directed links connect a participant with another participant he or she invited. If and when the number of participants become large enough, I will create a FB page to host all data and discussions on this project.





In the very likely event you did not read the post let me summarize. The prof wanted to study the mass blooming but the way he was collecting information was very inefficient in my view. I thought there should be a platform that eases the process; makes it easier for people to submit the data as well as for the prof to collect and study the data. Hence the inspiration for my project.  

Why fruiting trees? Well I looked out of my window and saw mango trees. 

Google Maps Engine

Today I received an invitation to take part in a project to document the sightings of Swallowtail moth (a species of moth) in Singapore. Click HERE to view the project. It is using the google maps engine.

I clicked the link and it brought me on my browser to a google map with many markers placed on it.
The concept is very similar to my own project, fruit tree mapping project.

I instantly created my own map and started exploring. So how does this google maps engine match up with the app I intent to create? Can I simply use this maps engine instead of building my app?

How would users generate the location of the sighting?
Unfortunately users would not be able to use their phone's GPS to pin down their location. However a search bar is available for users to search for their locations.

Can we differentiate between trees using the different icons?
While there are more than 100 icons to choose from, they would not be able to represent fruits very well. Custom icons can be used too but only if we subscribe to the service at USD5/month.

Can we attach photos of the tress to the markers?
Yes this feature is available. However you can only add photos using url or google search images. That would put off users from adding images.

Would users be able to search for and display only a specific type of tree (filter effect)?
Could not find a feature for filtering.

Can I view this from an app?
Yes I can easily create an app to open this map. An app would be beneficial as after installation users can launch the map easily instead of typing in the url into their browser.


Pros&Cons (Other than those already mentioned)

Pros: Data can be imported and exported using data sheet format.

Cons: Any user can delete/edit inputs of other. (This can be a pro too. It would be like the early days of wikipedia.)

Errors loading map

Our first aim was to make the map without the mapping function; the app will just load Google maps. Followed several guides to do it but when compiling ran into some errors which I could not resolve on my own. Am not sure which file to upload to Github to let Mahadir take a look at the errors.

Update on Github

Signed up for Github already. Have not created anything yet. First thing to do will be to create a ReadMe.

However it is not exactly like GoogleDocs. When we want to work on the code it will download a copy to our local system. When we are done we can upload it and then merge it. Any conflicts will be highlighted for us.

Decided to do modular programing in order to avoid conflict. Will try to get the algorithm and skeleton settled before Monday.

Working collectively

We realized that if we both work on the app separately it is not very efficient. Now we are going to look for a platform where we can work on the code together. Something like GoogleDocs but for Eclipse. Heard Github is a place for collaborative coding, we are going to check that out first. Currently I am traveling in the MRT but thanks to the Blogger app am able to update the blog. Very convenient.