Archive | May, 2014

Taarifa Hackathon Update and Upcoming

16 May

We camped out for a few lovely days at the inspiring Little Devices out of N51 at MIT. Hugs and hearts to Anna and Jose who brought their brains and their space to our hackdays. And didn’t find our requests for things like buckets, tubing, and CNC strange. We were constantly surrounded by go carts and craziness. So happy.


Mark used the OSM-Bright style and started to add Tanzanian Open Data to create a basemap for Taarifa (and generally anyone interested in Tanzania). Have started to refocus the cartography on public services as opposed to navigation – roads are the design focus of OSM Bright – Mark is changing this!

Requests have been made to import this data into OSM, however, as yet have been unsuccessful – I am following up on this! So, seeing as I can replicate OSM’s stack insofar providing a map tile service is concerned, I’ve started to create a Tanzanian base map.
Before: TZ Before


After: TZ after

Next Steps

Currently continuing to work on it. Will need to understand how the loading of OSM Stylesheets into TileMill can be re-consumed by others, currently the work is installed on a local TileMill and isn’t packagable for wider consumption. Changing this is a priority.

Lives at

On Mark’s computer – needs to be on a github repo. As a (rather large) side note. The finished product will need to be hosted on a server with at least 100gb of space, if not more. Consequently understanding where and who has this infrastructure will be important.

User Interface / Dashboard

Drew did a bunch of work on this. Here’s a drawing from brainstorming:

And a link to the PDFs.
Screen Shot 2014-05-16 at 7.24.56 AM

User Stories

It’s important to know who your users are, if you’re going to build any tool. By building out user stories based on experience and interaction, we can consider message format, connectivity, visual needs, and the like. Willow typed up a bunch of user stories based on her week in Tanzania.

Next Steps

There are some lingering questions that not enough experience was garnered in order to answer. Things like – if we want to show people that their reports are being acted on, but we can’t send them text messages back (if we’re protecting their identities), then having printouts at local municipal offices makes sense. What should those look like?

Code Base

Dirk and Florian led the charge on this.

  • Created initial installation and usage docs for the API and Waterpoints flavour of Taarifa.
    • Drew helped to guineapig this
    • range of bug fixes & doc fixes
  • Understand the sensor use case and how that relates to the Taarifa use case. Settle on the use of a separate, existing dedicated sensor api solution (potentially MITs chain-api).
  • Trying to complete the TaarifaWaterpoints application but ran into technical issues. Lots of debugging with Florian which lead to more a fundamental discussion as to what the scope of Taarifa should be.
  • Technical issues pretty much solved with a couple of things swept under the carpet for later. Scope discussion started on the mailinglist, community needs to comment (this is where a board could be useful)
  • Lots of discussions with Spencer (chain api), Jude (Promise Tracker), Ben (MoMo), Dan (Riffle) in between

Next Steps

  • Complete Waterpoint example so it’s usable
  • Main priority: focus on UI / Dashboard, implementing Drews mockups.
  • In parallel continue work (who?) on a more flexible, generic API depending on outcome of scope discussion on mailing list
  • Don’t actively evangalize Taarifa codebase to outsiders and other applications without some caveats.



Day 2 Updates

What Else?

There was also some killer work done on water sensors with Public Labs, and on telephony plugins with Tropo. Look for that followup here!

What’s Next?

Next, we’re taking this show on the road to University of Central London 5/24 and 25 (, and then to Dar es Salaam ( We hope you’ll join us, in person or remotely, for one or both.

See the list of our tasks on this hackpad page.

Updates from the Field 3

11 May

Thursday is a holiday, and so no meetings – we wake up early and head to Mkuranga District – a rural, rather than urban (like Tendale), slum. We run for the ferry, tho thankfully we don’t have to jump for it, new journalist friend Erin in tow. On the other side of the sea, we drive for hours, slipping between staring out the window and talking about interactions and plans. When we finally arrive, Msilikale talks with some women about if they’d be ok to be interviewed. We negotiate money amongst ourselves – in the US and Europe, you get paid for research subject time. Here, there’s an expectation that uzungu will provide money. I offer to buy a meal or drinks1, but it doesn’t go over. Even this is complicated, with potential larger ramifications. What expectations are we setting? Are those ok? Ethical? The lacking infrastructure and predictability isn’t just about drains and tap water, it’s also about social interaction and protocols2.

We talk about sewing, and water, and responsibility. There are only wells here, and those only produce salt water, with which they clean, wash, cook, and drink. There was once a community-held water point, but it broke at some point and it wasn’t fixed. The assumption is that the government will install the infrastructure, in the same breath as a complete lack of belief that it will ever happen3. Water can only be gotten when there’s electricity.4, when it can be gotten at all.

reports from the field

Again, there’s no central square, no place for known dissemination of information. Everything is done by word of mouth, neighbors talking to each other. Do they ever update each other with phones? No, there’s too much worry about the cost coming back to them (or to the person they contacted). But they’d be happy to do what’s needed to bring water to their place. If mgunzu like me try to figure out things, how can we avoid being jerks5? So long as the government brings it in, they’ll work with it. Again, this weird relationship to authority.

We hang out by one of the salt-water wells while Msilikale finds a taxi6, watching people bring carts and buckets to fill up. Children throw rocks at a lizard, chase each other, drink water from the bucket used to wash clothes now hung to dry. We stand under the gas station awning during a short but heavy rain, and then pile into a car for the long journey back to Dar es Salaam. Now it’s back to interaction at the scale of organizations, but now as informed as it can be on our short time scale by interactions with humans as humans, not in aggregate for logistics. The UNHCR refugee camp in Northwest Tanzania seems most appropriate for the water sensor innovation test deployment, as it’s a closed loop. Kibaha makes the most logistical sense for the test deployment of Taarifa, as a lot of cultural work around accountability has already been done there by a potential partner organization. Mkuranga doesn’t make sense because it’s too far out, there’s no existing social infrastructure for organizations, and there aren’t plans to put in water infrastructure for awhile yet, so people would quit reporting after awhile of no results. It’s all practical, but not cold. People here feel their responsibilities, just like anywhere else.

The next morning it’s raining as we gain a blessing from the Ministry of Water – we’ll work with their water engineers on updating reports of water points7. We sit in a taxi in traffic and talk, then meet with a potential local partner who will help with social interaction and embedding – managing expectations, closing feedback loops, continual interaction for a more successful launch – or for a better understanding of a failed launch. If it works in Kibaha, we’ll try it out in Mkuranga, with more focus on the sensors than on the reporting, to ease survey fatigue. We get back in the taxi and talk more while we head back to the Ministry of Water to talk to some engineers about what they would want out of a reporting system (yay more talking to people who use Taarifa, not just read the outputs!). As the depth of the water on the road increases, the speed of the traffic decreases. Finally, concerned about even making his flight, we send Mark off in the taxi with his luggage, and I pile into a bajaj with my own suitcase. A meeting to get to, and facilitate, on my own!

Everything goes beautifully. I’ve learned to hold firm when I’m told someone doesn’t have time, or tells me they only have a few minutes. “We’ll talk again on Monday, but right now I do want 15 minutes. That’s it.” Engineer B and I end up sharing frustrations, drawing on pieces of paper, and giving a firm handshake at the end. Msilikale and I meet up, and head to my new lodging – not the fancy hotel anymore, but a friend’s-of-a-friend house. From here, I can still see birds flocking, and the sun setting over the sea, but there are also bugs and the shower is weird and it feels so much more comfortable than the fanciness. We have dinner with one of her Dutch friends, and brave traffic, and bond over growing up in the Midwest. The ensuing days are similar, with one day blissfully off. Plans for Zanzabar are trumped by epic, amazing rains. I read frivolous articles on my iPad and watch the rain roll over the sea.

In all this, the World Bank8’s hammer is money, and so everything looks like a project to fund. What makes this a complicated mess to my anti-capitalistic heart is that, indeed, many projects do need funding in the current environment. I see the “we read as much about about a grassroots thing that works as we could, and this is how we think we should do it…” All the people I’ve met within the org want a way to make the world to suck less. But these are institutions whose tools are people, and funding, and other institutions. And while they try various tactics, and sometimes make headway, in making the world suck less… they’re also held accountable for their actionsIn theory.. The difference is, the people in grassroots initiatives have to live with the reality of the failings and successes of their (and institutional) endeavors. So of course they are who I think of first. And last. And always9.

And Msilikale and I go over the drawings I did, and listen to music, and talk about all sorts of things. The power goes out, and we keep talking, the windows closed against mosquitos and the oppressiveness of the growing heat inside overwhelming. We walk in the dark to a local Indian joint, eating overly peppered food and listening to the calls to prayer out the window. The lights go out there, too, and we eat by the light of cell phones until the generators kick in. “This,” say Msilikale, “is Dar es Salaam.”

1. Worked for Kibera.
2. Scott’s Seeing Like a State is ideologically interesting, but if there’s no way to get clean water but through organized distribution of resources, such ideology gets tempered at least a bit.
3. It’s like breaking up with someone before they break up with you.
4. Mind you, this is a project with the Ministry of Water. Not Ministry of Power. This is with water. So we can only focus on water. *shakes fist at silos*.
5. Again, Msilikale mitigating anything that seems like a promise. Or hope, really.
6. Easier to negotiate price if we’re not there.
7. The hand washing tap in the MoW does not in fact produce water. Oh, the irony.
8. A World Bank innovation fund is what is supporting this initiative.
9. Not saying others don’t, simply that there sure does seem to be a lot of reminding.

Updates from the Field 2

10 May

People don’t lean out of vehicles to ask for directions here, as they did in Nairobi. The security guards, when they exist, lounge in chairs and ask questions rather than standing, mostly silent, with automatic weapons. But like Nairobi, meetings can start (and end) hours late due to traffic, to tardy risers, to rain that causes traffic, to conversations running long, to torn-up infrastructure (that causes traffic), to slow service for your lunch meeting. Tardiness is sometimes used as a sort of posturing card to play – whether or not someone gives you a meeting, and how prompt they are, as sorts of indicators of status.

Everything we’ve done here has required status strutting in order to gain speed. Those plates I told you about? Only people with yellow plates get pulled at traffic stops, because they won’t be diplomats, or military, or donors, or government. If you get stopped, it’s easier for everyone if you just hand the officer money rather than pay the huge fine for driving illegally. Which most are. And because we don’t want to slow down our project, nearly every introduction is “and this is Willow, from MIT.” Which is great and all1, but as someone who prefers to be affiliated with institutions for access to incredible brains and the space to consider at length, rather than constructed legitimacy, it makes me feel like I’d be prettier if I just smiled2.

Those same posturings and rerouting the system means there’s a fear of transparency here. Entire systems here are built up around being opaque. People across all walks of life ignore the floods for fear of being blamed for what happened, being put out of a job. This means any transition into transparency will require safe space. No “we’re coming after you” attitude, but a “we have been operating to the best of our abilities within the structure we have. Now that the structure is changing, we get to change as well!” But it seems enough people into open data and transparency have done it with a vindictive streak that everyone balks these days, and it’s a slow, gentle process.

Two and a half days into an 8 day trip, we’d chatted with NGOs, World Bank3 employees, the Ministry of Water, UNHCR, and no people who actually live the experience this technology would change. And as much as I trust all of those people (and my hosts) to know what the status of their work is, I needed to go see things. Establish ground truth. Everyone gets caught up in paperwork, rhetoric, image, and email, and so seeing it for oneself is always imperative. Mark, an amazing guide and cohort as always, got us out to Tandale a full day early for the sake of my patience and sanity.

On the way there, Msilikale leaned forward to ask the bajaj driver to drive like he does, not like we’re tourists. The roads had washed out from the recent floods, full of pot holes and rubble to negotiate and lurch over. Tandale is a slum in Dar es Salaam, and is a place to live, just like anywhere else. We walk with a man who has lived there, him taking us past houses with water lines up to our mid-thigh, insides still covered in silt, to the river running by the open defication area (ODA here). Kids run by with make-shift toys, and young women peep out to ogle my hair (or Mark’s size)4. As we stand by a washed-out bridge, our guide explains context.

The river divides two areas, one has most of the markets and the other is mostly houses. There’s no grid system (it’s an informal settlement), and so paths are highly reliant upon available bridges, and new structures are based on those paths. IE, functionally ad hoc labyrinthine. And when the floods came, the bridge washed out, and there are still people learning about that and figuring out new routes home. No one is responsible for the bridge – the government won’t come fix it, and the man who built it and had lived nearby died awhile back. No one takes responsibility, either.
It’s not just the bridge getting washed out – it’s the height at which the water rose, and that the ODA is not much higher than the river on a regular basis anyway, and it’s certainly lower than the water marks. So all the trash and bodily functions and such from the ODA get caught up in the river, which means it gets clogged up (as do the drains in the area), which means standing water, which means more mosquitos, and mosquitos are bad news bears. That’s besides all the things in the ODA also flooding into the houses with the rest of the river water.
Flooded or not, the water from the river (and the wells around it) is used to wash, cook, and sometimes drink. This is only for people who can’t afford the water out of the water points – which can be salt or fresh. For half price, you can get jugs of salt water to wash with, and sometimes for cooking. For full price, you can get fresh water. The water is delivered out of water points – giant containers, raised up off the ground, from which people purchase water. These are privately owned, though installed and supplied by the government, and few enough are stocked and working at any given point that there are queues for the resource.

I ask questions. How do people know that one is broken? I wonder about where knowledge of what is going on would be hosted and shared. Is there a town center where things are posted, or is it all word of mouth? No center, just neighbors telling neighbors. No bulletin board system. I saw a vanishingly small number of feature phones while we were out. Only government and donor officials and contractors smart phones. Which, to me, makes me wonder how to get maps built of the water system back to the people in the area5. And given the level of corruption in the country, that data being only accessible to groups already in power is fraught with peril. Mark, aware to not ask questions about anything that won’t have a guaranteed resource response, waves me off asking more specific questions of how points break, and how to track response, what would next challenges and steps be, etc6.

Night threatening and malaria mosquitos7 lazing about, we walk back towards the main road. At a bodaboda station, we negotiate with the fliers. I feel like I’m choosing who to be auctioned off to, having to dismiss the enthusiastic and be wary of the most aloof. Msilikale shoos off the most invasive, and I’m glad of a native Swahili speaker friend and a friend who’s 6’6”. That negotiated, we take a rather epic dive-bomb through traffic route home8, avoiding jams and vehicles traversing the median to gain a clearer route. Only one or two of the passing bike riders make kissy faces and eyebrows at me. Mosquitos die on my visor, and mud splashes on my boots.

And I stare at the bathtub in the hotel room, and think about not finishing my dinner when I was young and knowing most parents referred to starving children in Africa. But the issue then, and with this, is in supply chains and politics around them.

1. Waves MIT flag.
2. Institutional objectification. Which is at least different from institutionalized -isms.
3. Yes, of course I wore my “/capitalism” pin, why do you ask?
4. Please forgive my tense-shifting.
5. A core ethic when obtaining data. See also this blog entry.
7. Dhengi for day, malaria for night.
8. Sorry, mom.
9. Inside joke.

Updates from the Field

9 May

6.5 hours from JFK to AMS, and another 11 from there to DAR. Woobly from hours on planes, binging on movies, and clandestine email response; I stood in a pen full of anxious people waiting to regain their passports. I watched the processing, detecting patterns but not defined process – most passports and paperwork went in one window, in a pile, often added to the bottom of a stack but not always. Person there did something, often interrupted, passed on passports in single or in aggregate, not in the same order received. Then passed on to one of 3 other people, who did… something else. Of the two windows which kicked the passports back out, one would use a mic and announce your name, the other just held up the passport and people would see their image (or not). Those unclaimed, plus.. other ones? to be distributed were carried through the pen of passport-plebes1 with names shouted or mumbled. Finally escaped, Mark and Rav met me past customs, and we crammed into a car with a misfit axel, grinding gears though city streets. Traffic lights were blatantly disregarded, motorcycles passed on either side, and we attempted conversation over the loud and the heat. We got to the hotel, one of the few that’s approved for World Bank staff2 to stay at in Dar es Salaam. I showered the 20+ hours of travel time off, well aware that many people in Dar don’t have access to water3.

Monday I headed out to meet Mark at COSTECH’s innovation space4. Could I walk? Ha ha, no. Was there public transit? Not worth mentioning. So another cab ride, the driver and I talking about corruption, and family, and why he loves living in Tanzania. There’s no war, it’s peaceful, he doesn’t worry regularly about being killed. “How long has it been peaceful?” I ask. “Seems like from independence.” “50 years?” “Our independence was in 1961, so 53.” “Congratulations.” “Thank you.” He rolls up the windows when we come to lights, pointing out people on the side of the road, says they will try to take my watch or bag or phone, because I am mzungu. Do I know what that means? “White bread?” He laughs. “Your hair, though, it is blue.”

At COSTECH, Mark and I chat with other people. The local developers who worked on Taarifa a few years ago, have continued to develop and map. A Fin from TANZANICT joined us, and Mark talked through the ecosystem of water projects5, my drawing furiously to keep up. From there, we hopped in a bajaj speak to a large NGO which has been in the area for a few decades. During the ride over, Mark points out license plate colors – blue for diplomat, yellow for private cars, white for public – my dark humor latching onto the hierarchies embedded in such a visibly manifest way. It makes me want to actively avoid the shortcut of institutional credentialing.. but we don’t have time to not take them. At the NGO, we sit for awhile doing email, the African-pacing of time reminding me of Rob’s laptop sticker and conversation I went to Kenya – There is no Hurry in Africa. I drink sweet coffee we chat about Swahili having at least 3 ways of saying “I’m sorry,” my suggesting that and the side of the road driven on as main indicators of English colonization.

We finally chat with the folk at the NGO – for hours. Both groups circling the other, Mark being performative in his role with World Bank, Rav as backup in stitching things together, myself trying to pick up on social cues and attempting to not speak too quickly. We talk about accountability, transparency, scaling, and survey fatigue. If we ask people, again, to provide information, what do they get in return? So many maps have been built, so many initiatives have blown through, and life still sucks. What we possibly do that is any different? Can we work with the local municipalities and national water ministry to enforce the fixing of the points? We’re working on it. Can we make the information visible to the people who live somewhere, provide material and structure to advocate for themselves? That’s a long and difficult journey, but possible. We circle each other for awhile, uncertain of if the other party “gets it,” from the social responsibility or the data possibility sides. Finally seeing that we do, we agree to send a draft MOU, and we head out in another bajaj. This time with Mark, Rav, and myself.

Now, these things are tiny, just big enough for two people plus some wiggle room. Here we have 3 of us, of whom one is over 6’6”. In the strange layering of apologizing, stubbornness (from all parties), and negotiation of money, we make our way back into town to have dinner and pile through emails. We walk home (hooray!), Mark stating time and time again “not a tourist” in Swahili. We get freshened up and head out to see a friend. As it’s rush hour, we pile onto a bodaboda, or a motorcycle taxi. Both of us. Making three people. We ride like this for awhile, Mark asking passing motos if they are also bodabodas, offloading onto an available one, us easing between lanes of traffic and narrowly avoiding potholes. Sometimes we ride on the sidewalk. When we arrive, we drink beer on a balcony, talking about teaching coding and entrepreneurship, discovering what patterns work across places and what must be thrown out. I find a difficult conflict in myself, between a growing awareness of levels of corruption, and my deep need to defer to people who live the reality of this place. I think back to conversations I had with Lorraine over lunch at Theorizing the Web, about how people are able to use any system to still do good, and you disrupt them as well when you shift systems. And then a car ride to dinner with assessors of programs. Amazing Indian food, and conversations around baselines and statistically predictable incongruities, and how to learn from things even as you fail from them. And beer in a place called Cuba, which we joked I couldn’t get into. And then finally back to the hotel to sleep.

The people in the local offices seeing people like me, who are just coming in for a short period of time, like some sort of Starship Troopers, shouting about how someone else fucked up while things they don’t understand happen. I try, as always, to provide scaffolding for others to see things in new ways, rather than complete deferment or frustrated attempted mandates. The local groups here are doing incredible work, and it reminds me of spanning mutual aid and specialized response. Here, in practice, are many things I spend brain cycles on – philanthropy as unsustainable, colonialism and aid, organic discovery and institutionalized knowledge, and digital divides.

1. Done more for alliteration than social commentary. Yes, I realize how self-referential and socially (un)aware this comment seems to be.
2. Not the IMF. The World Bank. Mark does good work there, and it’s who has contracted me for this trip as well.
3. Cognitive dissonance jazz hands.
4. Which I needed to get to, on my own, without data, in a place I’d never been. I sat with my anxiety over loss of control, of the possibility of getting lost, and hugged that part of myself to acknowledge it.
5. As always, a mirror for my own reflection as well. The pace at which Mark moves, trying to loop people into the understanding of a complex system of technology, people, and politics in his head made me think about what I expect of people and how I express those expectations. It’s like when I speak about either of my friends called “case” – one of which is spelled that way, the other of which is spelled “qais,” and the difference between them is so clear in my self that I don’t think to differentiate them for the person I’m speaking to.

Overview Video

8 May

Taarifa Hacks Series

8 May

We’d like to invite you to come hack on data, software (front and back end), hardware, and all the bits between in Cambridge at MIT’s Little Devices on May 7th and 8th (, in London May 24th and 25th (, and in Dar es Salaam ( May 31st and June 1st.

The Challenge

There’s not much in the way of access to clean water in Tanzania. In the informal settlements, there are a bunch of water points, but many of them are broken. Rather than a continual process of putting in new ones, the local water engineers want to fix the existing ones – but they don’t know where the broken points are. This also prevents large-scale response organizations from accurately deploying resources (and seeing what initiatives are already working).

Our Approach

Through a combination of participatory mapping across a few groups, and water sensors, we think this situation can be bettered. The incoming information would not only feed into the repair cycle and communal awareness, but also into larger governance decisions.



The Software

Taarifa is an open source web application for information collection, visualization and interactive mapping. It allows people to collect and share their own stories using various mediums such as SMS, Web Forms, Email or Twitter, and then the stories are placed as reports into a workflow. At the event, we will be building out its capabilities and localizing it to Dar es Salaam.

Things we might work on with the software:

  • documentation (tutorial, installation walkthrough, waterpoint demo)
  • general testing (unit, integration, ..)
  • modern UI / dashboard design and development (UX, d3, ..)
  • dummy data generator
  • pretty report generator
  • SMS reporting
  • demo phone reporting app
  • access control & user management (what is needed? how to best implement?)
  • explore relevant standards (e.g., open311) and decide what is useful to support
  • auction model for work/support contracts
  • how to subscribe and notify people of changes to ‘their’ report (Twilio useful?)
  • how to integrate sensor readings? data model?
  • how to deal with report attachments (images, videos, …) currently not supported
  • nice continuous integration setup & auto deploy to heroku when pushing to master
  • exploit linked data for reports, locations, people, and resources where possible

Would also be useful if we had a second use case (besides waterpoints) to double check assumptions and architectures and test the dev experience. Promise tracker?

The Hardware

We’ll be playing with the Riffle (Remote, Independent, and Friendly Field Logger Electronics), a low-cost, open source hardware device that will measure some of the most common water quality parameters, using a design that makes it possible for anyone to build, modify, and deploy water quality sensors in their own neighborhood.

Why This is Different

I tend to be wary of social good hackathons. At Geeks Without Bounds, we run 1 to 5 internationally every month. Many of the projects are a learning experimentation for the people who attend – which is awesome and worthwhile. Few social good hackathons are for production, however. This one is. Taarifa is already deployed in Ghana and Uganda, and we have a place to deploy in Tanzania. People will use this tool: the few people with feature phones and connectivity, the water engineers and camp staff, and aid organizations.

Where you Come in

We need your skills, questions, and energy. The project is clearly defined, as are the times to work on it. We want you to join us at an event if possible, remotely if you can’t make it in person, and to continue to be a part of the community even after this round of open-source hacking.

We’ll have a call-in this Sunday at 19:00 EAT (worldtimebuddy for conversion). Bring your questions about the project, what you need to prepare for the hack weekends, or just out of general curiosity. Register for the hangout here.