Information and Communication Technologies for Development

We have made progress on the individual parts of the solution of collecting data from mobiles and sending them over and storing the forms in a database for query and reporting purposes. We are meeting on Tuesday to summarize the project status and create a few slides in the process.

Stay tuned

Saajan wrote an interesting feature story about our project and his trip to India in The Tech today. It showed up on the front page. Check it out:
http://www-tech.mit.edu/V128/N18/india.html

We're now working on a proposal to solidify our plans for the future. After Saajan's trip to India, we have a number of promising possibilities. I won't talk about them on the blog yet. But once we get things finalized, we'll report back. Technical development is currently stalled until we finish up this proposal, at which point we'll focus our efforts on some design meetings.

As the project progresses, I seem to gain more exposure to the realm of disasters and the role of communication. On the one hand, there is ICT which helps to promote rapid dissemmination of information. On the other hand, I think of the role of how "communication" in general can help to solve problems. For instance, going off of our group discussion on Tuesday, imagine if certain information such as that on possible floods or cyclones, was communicated with citizens as soon as it was gotten. Perhaps a lot more damage will be mitigated.

Here's our midterm presentation:
We hope to have some of these slides available in text/image form on our wiki as well.

| View

We've been making great strides over Spring Break on the back-end system. I'm starting to get comfortable with the development environment. Our back-end code base has been moved over completely to .NET, and we now have a programmer who understands the basics of the language.

However, as one challenge passes, another presents itself. We have described a high-level overview of the system as a whole, but we have not gone into much detail regarding how specific parts should behave.

Here are some questions we need to ponder:
1. Our system must have the ability to access data quickly and organize it effectively. How will the data we receive be organized?
By location? By "cases," like specific disasters? By tags? What metadata should we standardize across the system? I really like the idea of storing pieces of data (i.e. pictures, filled-out forms, voice memos, etc.) as "records," and then tagging them with appropriate information, like location, case, etc.
2. Saajan's reports on the wiki have expressed a need for an early warning system for floods. How will we integrate this into the system?
3. How will cell phone users find the right forms for their needs?

Muzaffarpur (pron. Mujaffarpur!) , Monday March 24th - Niraj and I left Patna at 6:30 this morning. The journey was eventful - on the outskirts of the city we came across an autorickshaw which had toppled on to its side, trapping the driver underneath it. We stopped to help right it, and thankfully the driver wasn't badly hurt, just bruised and shaken, so we carried on, reaching Muzaffarpur around 8:30.

I started the day with a meeting with Nitish, who's in charge of the Muzaffarpur office, and Niraj. The rest of the day was spent going to two villages in Musahiri block, which is an hour or so away from Muzaffarpur. As before, I've put notes up on the wiki so I'll just pull out the things which I think are particularly interesting or important.

The common theme that I heard from Nitish and from the villagers I met today was the need for some sort of early warning system for floods. This is not really something which came across in our discussions with CRS Delhi, but it seems like it would be extremely useful. It would probably take the form of
villages which see a rise in water levels telephoning a control centre which would analyze this information, perhaps together with weather reports and satellite
maps, and issue warnings as appropriate. The Inter-Agency Group (IAG), a state government-organized conference of NGOs is working on this right now, and I've asked for some more information on the direction they're taking.

Patna, Sunday March 23rd - Unfortunately, due to it being Easter Sunday and the day after Holi, CRS couldn't arrange any meetings for me today. I'm off to Muzaffarpur at 6:30am tomorrow and should be meeting with some of CRS's partners who are working on the water supply in the villages. Luckily I typed up the water sanitation survey in XForms before I left so I don't have too much work to do to make a useful demo to show them. I'm hoping to meet the AirTel engineers on Thursday - I should be back here early in the morning and my flight's not until 7:30pm.

Patna, Saturday 22nd March - I met with Jennifer and Bipul from CRS Delhi this morning. My notes on the meeting are up on the wiki, so I'll just highlight a couple of the more interesting and important points.

It seems like our impression of the problem scope, at least as far as management in Delhi is concerned, was pretty accurate. Jennifer most definitely does not
favour a .NET/Windows Mobile solution; one of the first things she said to me when she saw Arvind's HTC phone on the table was that she was not keen on
buying them for all of their staff, so it seems like our gameplan of client-side Java + server-side ASP.NET is a good one. They asked about the possibility of tech support from MIT - we'll have to discuss this but I'm sure we can work something out, at least for the first year or so.

They do see this as the spearhead of a large-scale deployment of ICT throughout the organization, so we have to get it right. They have a few ideas for other things that might be useful, which I'll discuss further when I'm back in Delhi on Friday and I've had a chance to actually look around in Muzaffarpur.

Note: this post is entirely off-topic, stay tuned for a relevant post!

Patna, Saturday 22nd March - Whenever I travel long distances, especially when I arrive in the middle of the night, it always takes me a while to properly realize that I've arrived. In America the thing which really drove it in was (no kidding) seeing a 7-11 for the first time. In England it's the way that whenever I land in London, no matter what time of year, the temperature is ten degrees Celsius with light ain. In India it was Holi, the festival of colours.

I'm blogging from the departure lounge of JFK - how often do you get to say that! I should be arriving in India around lunchtime tomorrow Boston time, where the weather is a cool 25 degrees (thats in the seventies for all you Americans...)

The plan right now is for a preliminary meeting with the staff in Delhi on Saturday. I'm flying out to Patna Saturday afternoon, and then on to Muzaffarpur (where the Bihar regional office is based) on Sunday morning. This is somewhat subject to change as the Holi celebration (the Hindu festival of colours) is taking place on Sunday in Bihar, as well as it being Easter Sunday, so I will just have to play it by ear - but that's pretty much what I expect from India anyway! More on my plans for the trip below the line.

Luis's gentle but stern comments on the very rushed paper report gave us all a shot in the arm.

We discussed earlier our solution for helping CRS India with their data collection needs during disasters and during other times. We decided upon two main interfaces for the system: a web interface which could be used to enter form data and view submitted data, and a mobile interface. For our initial prototypes, we've been focusing on entering, not viewing, data on the mobile platform.

I'm responsible for the system's web interface and back-end. Students as we are, both Saajan and I have mostly had experience with open-source development tools, since they are free. In our initial discussions, we assumed we'd be developing the back-end in PHP and MySQL.
More after the jump (including a flashy diagram!).

I currently have four cellphones (plus my own) strewn out over my desk for debugging purposes (as well as a GSM modem plugged into my computer). I hope all those scare stories about cellphones irradiating your brain are not true!

I am beginning to think that the most time-consuming part of the cellphone app development will be testing it on all the devices we plan to deploy on. CRS are keen to use Windows Mobile devices, so that's a key target platform (which is interesting because that seems to be the only platform which currently supports CLDP1.1).

We'd also like to see it running on lower-end phones so that at least a limited subset of the functionality can be deployed to many more people. One of the issues they have at the moment is that while many of their on-the-ground agents have access to Internet-capable PDAs and laptops, they are often some distance away from where they are needed. If we can deploy our app on a $30 cellphone, it will be economical to have someone in each village who can rapidly report to CRS as requrired.

After delays in getting one component or another, finally things seemed to come together today. Still needed to work on the travel plans for Saajan. Java code coming together for the demo planned to be ready for the visit. Banke and Arvind getting a couple of surveys together to ask potential users of the system what they felt were important.

As usual nothing is ever entirely straightforward with Linux, so here are a couple of tips if you're struggling.

Graduate student in the System Design and Management program which is a joint program between the MIT Engineering Systems Division and the Sloan School of Management. I have 12 years of experience in communications technologies (landline and mobile). I have a previous Ph.D. in Electrical Engineering from Temple University Philadelphia. I am excited to be able to make a difference in disaster relief measures in my home country. I would love to be able to design effective solutions to real problems.

We have started getting together and functioning as a well oiled group. We had a few discussions on how to do and what to do. We have concluded that nothing can elicit more reaction than a mock up of the mobile application that would enable a form style entry of disaster information. We are leaning towards me going over during Spring break time to shake out the requirements and meet with all the primary stakeholders including the boots that meet the ground when disaster strikes. We will be most likely talking to CRS India folks Wednesday evening EST, more info after that meeting. We have to get to final deployment before July 2008 which is when the monsoon hits. Banke will be working on the survey design and who all we want to talk to. Nick and Saajan are our primary technologists and are currently evaluating Java/PHP/ HTTP vs SMS options and even a potential FAX solution when all else fails.