A week in Tanzania

Recently, I had an opportunity to work with the amazing team at School of Data, meet interesting people and do cool things around data- thrills me. The World Bank and the Open Knowledge are working with the Tanzania government on her open data initiative. We spent a week in Dar-es-Salaam, the largest and richest city in Tanzania.

Last week I worked with the Ministry of Water and National Bureau of Statistics of Tanzania. The goal? Teaching data skills and kicking off an Open Data Initiative for the government of the United Republic of Tanzania. We made a three-man team: Michael from School of Data and David from Code for Africa. Cool stuff going on in Tanzania! The most interesting thing about our work there was the idea of not knowing what to expect, in terms of data to work with, who does what in the ministry, how all the departments are interlinked and what skill level to expect- finding answers to these questions became part of the one week training.

Deep dive

And so we made a deep dive into the unknown waters of Tanzania water data. First we tried to figure out who was in the room, which department they were from, what their role was in the data work-flow process and what data they usually work with. We also had a discussion on the the data work-flow process in each of the departments represented in the room. One exercise turned out to be so telling: working in small groups, the participants were tasked to randomly list the type of data they usually work with. Next, the list was pasted on a wall giving off a mixed-up list of stickies with different categories of data. The last bit was to categorize the list so that we had unique groups/clusters of types of data. Super cool way to get people talking, become ‘data conscious’ and learning about each other’s work!

Image

Tools, tools, tools

The next two days were about exploring tools that could be used to work with data. There’s so much you could do with data, but there’s usually the first things you do with most datasets. Find out if it’s good for use: say analysis, visualization, building models or merging with other datasets. It it does not fit the bill, then it probably needs some data cleaning love. This was the case with some of the Tanzania datasets- missing values, duplicates, unavailable meta-data to give context and most likely a need for data quality consciousness- what passes as good quality data, what are the not so good experiences with data quality and some best practices for data quality management.

We worked with OpenRefine which is an open source tool for cleaning, analyzing and plotting data. There’s so many ghosts that open refine will unearth about your data. For example names that may refer to the same thing but are spelled slightly differently, say Dar es Salaam and Dar es Salam. This is the same place, but you want quality data? Then think consistence. We had spreadsheets as a compliment.

Image

We also took a look at the water geodata, creating maps and running queries to answer questions like where are the functional, non-functional water  points? What is the ratio of functional water points to the population served, which of the water points in the Dar es Salaam region are handpumps among other GIS exercises like styling, coordinate reference systems and data conversions, why use .txt files instead of .csv files when making a map. After giving all the data cleaning/mapping love, we thought it a good idea to introduce data visualisation with google fusion tables.David from Code for Africa did an amazing job with this.

And then we hosted a water data expedition on the last day.

Smelling the flowers?

Dar es Salaam is a beautiful city. This was the view at breakfast- a harbor! The abode of peace indeed! Oh, did you spot the national flag of this United Republic….nicely hoisted on the pylon, in the center of this picture.

Image

Aside

Data beautification, some ‘prototypes’

Recently, as we’ve been doing lately, had the Outbox Hub host us,  to talk about data visualization but also get our hands dirty creating data visualizations with the most unlikely tools- paper, straws, pens, bottle-tops, the floor and wall! The design inclined refer to it as designing data or making it beautiful. So we took a deep dive in.

We organized this mainly to discuss best practices, highlight worst practices from a design standpoint when creating data visualisations. The main goal here is to have your data communicate, clearly to your audience-it could be to children, teenagers or adults.

Take a look at some of the varying interpretations of the data:

SAM_2297

Raw dataviz from the meet-up!

From the data design experts:

We had a great mix of design, communications and programmer experts to share their insight, with 25 people in the room: Joe Ssekkono a front end developer and visual artist gave a presentation explaining ‘what data visualisation is’ , ‘how to plan your data visualisation project’ and ‘evaluating and iterating your data visualisation project’. He went on to show us a step-by-step process of how to start data visualization, some tips & tricks, websites and handy software tools. You can find the complete presentation here.

Emmy Van Kleef a Creative Communications and Concepts expert talked about the ‘latest trends and available tools’, the ‘best practices and worst practices of data visualisation’.

ImageAdapted from Emmy Van Kleef’s Presentation

Eve Ndagire a graphics designer talked about what ‘categories of data you could visualize’ (is it linear, discrete, continuous or categorized/numerical)? There are several ways of visualizing data, for example using bar charts, histograms, graphs, maps, pie charts, scatter plots and sometimes very abstract methods. We decided to go traditional and practical, using non digital tools like paper, pen, straws, crayons, bottle-tops, the wall, table and floor.

The workspace was very busy as everybody worked away, trying to visualise crime and number of convicts in the major cities in Uganda. A break did not seem welcome to the participants. Take a look at the dataset we worked with.

ImageNext-up

This turned out to be an interesting topic, and there was demand for a similar meet-up with more emphasis on using digital tools/code/software to make data beautiful. We hope to explore some of the trending data visualization libraries in the future. This meet-up allowed us to create ‘prototypes’ which we’ll build on, the next time!