When Real-Time Location Based Data Makes a Difference – Saves Lives

August 27, 2009

in Knowledge Innovation, Mobility, research, Wireless

I’m curious. Are you ready to contribute your mobile to the common good? Would you participate in surveys on an ongoing basis? If we made it easy? Would you join a panel? I’ve been thinking more down this path after my visit to the Amazon.

In the Amazon river basin, Dengue fever kills. Each year, as the floods reside, the mosquitoes breed and epidemics break out. There is no vaccine, and if you are lucky enough to survive you may have lived through 40 days of hell. This year  there was no huge outbreak (factually cases in this region declined from 3500+ to less than 250) and any reports of Dengue were quickly reported and contaminated water treated. See my earlier write-up here. What worked the miracle?

Other than a Nokia mobile phone and a little survey software nothing else changed. The same health workers were in the field. The same samples were dispatched to the lab in the city. However this year, each report they made was done on a location aware mobile device which instantly (timely!) forwarded the data to the central processing facility. This eliminated response time – to get data in / re suspected locations, errors in writing and submission, and lost of data in transit. In Manaus a simple little survey program made life easier for data-capture, submission, and response.

Back at INdT (Instituto Nokia de Technologia) I sat down with Andre Erthal who leads the Community Development Group to understand the Data Gathering software and how it all worked. I’d already been impressed with how easy it was to input the survey data in the field. Yet making it all work requires talent and understanding of usability, connectivity, and useful output.  As Andre explained it, he left me wanting access to their software.

Observations of specific areas important to Mobile Data Gathering:

  • The survey must work effectively on a mobile phone
  • GPS enables results to be mapped accurately
  • Each phone has a unique identifier
  • Survey updates must be tightly controlled, synchronized.
  • Various data-upload options must be provided (particularly for emerging markets)
  • Must be easy to prototype questionnaires. The use of a simulator is a bonus.
  • Easy for the field group to use and enter data.

As Andre explained, they built the Nokia Data Gathering system in Adobe Flex and all core data is in XML which makes it easy to read and later manipulate. He began by demonstrating how one creates a new survey. The tool was running on a Windows PC (soon to be ported to the web). In practice, survey creation isn’t much different to any online web system. It’s easy to create categories for questions and then use different formats, text, numerical multi-choice etc. In practice there is a minor psychological issue in recognizing you are creating a survey for a mobile. Thus the situational input needs taking into account. There’s also the option to use skip logic functionality. Once the survey is complete you can run it on a simulator.

The tested survey was then uploaded to the distribution server. It was here that a key point was made. Once you have uploaded it you cannot just write over it again. The key is to keep the survey synched across Mobile devices and as they may be offline or some may have already responded, questionnaire changes while a survey is in progress, isn’t a bright idea. Going live may also happen at a time different to “loading the survey” when a device may be  online. Survey results are tied to a phone’s IMEI as the SIM may even change from time to time. Obviously GPS coordinates are provided each time a questionnaire is submitted. More advanced deployment will come with time.

I was impressed by the focus here on Data-Gathering. This is really a technology to empower other organizations and governments to capture real-time location based data. The implications for other services beyond the health care tests done in Manaus, Brazil is restricted by limited distribution of the tools so far. The benefit is the instantaneous capability to see results on a map with a high degree of accuracy. Many organizations that collect data currently on paper can move rapidly to mobile. The benefits include:

  • one-time data input at source – no re inputting of data / reduction in input errors / time savings
  • accurate location based reporting and instant result mapping
  • consistent structured data formats XML
  • easy deployment and real-time collection from the field – reduced delay

In Manaus the mobile data teams were using GPRS/data connections to collect the results. Yet the program has been optimized for third world usage where data connections may not be available (GPRS was the cheapest input option in the Manaus area). So surveys may also be submitted via SMS. An SMS submission rather than using HTTP converts the XML data into a binary SMS format and is then sent using multiple SMS messages. When received they are converted back into the XML file. I’d think in some markets this would be much more expensive for data submission. Yet effectively we have SMS almost everywhere.  So the data gathering program provides real opportunities in the most undeveloped markets. If the SMS fails, the data is saved to the memory card which can later be uploaded. There are options for security eg https and it works currently on Nokia devices like the E61 with full keyboard and long battery life.

Of course I also wanted more. Will it upload pictures or video? Can it play product examples? All things that can be built in and I’m assured will be appearing in future versions.

This left me with more questions.

I kept asking…. why can’t I sign up  for a beta? I’m still being difficult about this. I’m not sure I understand what the market model or service model is for this program. Nokia says they are using Open Source tools etc and are planning to release it for non-profit and governmental organizations.

Collecting data is potentially an empowering technology. The benefits exist when software like this can be rolled out rapidly. My suggestions:

  • Open it up as a beta take the names and suggested projects. Let users host on their own servers  and simply experiment with it.
  • I’d release it under a license that is similar  to WordPress. Eg let me host my own system and if I want run / distribute many questionnaires. If it really is a business then a hosted service is always a possibility.
  • I can see this being quickly integrated with a WordPress installation or similar.
  • Survey tools really aren’t that difficult. This one will be copied and there are others already out there. I found this to be elegant although I’d want deployment on a broader number of phones.
  • I’d also like to wake people up…with a message or notification. This may be a survey or something they have to gather or report on. Eg have latent researchers in the field. Or have the opportunity to interview participants based on where they are.
  • How can we learn faster – 1) more questionnaires and different types of problem sets. I like that the focus of this is “data capture” factual data / reporting where speed is enhanced. The interpretation has been left to the organizations gathering the data. That’s smart.

Most importantly this project demonstrates that we have to think about data entry now in terms of mobility, location and real-time. The cost of this enhancement is trending to zero.  There will be huge benefits in the developing world and yet we would be foolish to ignore them locally too. Some examples that make sense:

  • Delivery Services: The third world trucking co/group can use this same type of data process to keep track of deliveries at all times. Infrastructure cost almost zero. FexEx in a few mobile phones.
  • Agriculture: Can keep track of crops, pests, disease, and provide “notification” “zone services in response to participation.
  • Villages / Cities can execute accurate survey assessments and conditions. Locate environmental problems or simply use for census data,
  • Emergency Relief: When that next disaster strikes there may be nothing better than ground reports, and estimating how and where supplies are needed etc.

What other ways could Nokia harness Mobile Data Gathering Tools?

It might be interesting to put a questionnaire on every new Nokia phone sold in emerging markets. It won’t work for everyone but each one would provide….. a location of where it is first started up or at a specific run time from new and perhaps capture some details about the user. I’d think little clusters of data might just show where the next wave of phones are reaching the unconnected… More importantly, Nokia could show how providing a little information now and then could really improve lives.

When it becomes easy or almost routine to harness mobile data gathering who creates the mashup will change. Currently the best map mashups are things like real estate for sale on a map. These use relatively static data sets that were seldom based on real-time GPS data sets. I think there is also a flip side to this. It’s simply advertising what you need or have to offer. A simple location based classified.

Data Gathering is just the beginning. The data explosion is coming. We won’t only advertise we will also broadcast other data or monitor other things. Eg I may choose to monitor air pollution today. Of course there’s nothing then that I may have to do… to provide data I just move around.

  • Stuart – this is a great write-up, and quite inspiring! I also like that it’s an open source tool, with potential application for NGO’s. At the info-activism camp I was a part of in Feb, there were a few folks risking their lives to bring about political or socio-economic change in countries ridden by repressive regimes. Would something like this help them getting word out on say, atrocities, irregularities etc. without risking their lives?

    As always, you’ve shared future possibilities around the Data Gathering Tool, and it’s making me think of how it might apply to other emerging markets, for applications outside of health care, and more personally as a survey/data gathering tool in our work as researchers.

Previous post:

Next post: