This post is written by Habtamu Woldeamanuel Habtemariam, Database Manager on FTFE VCA

During KNOWvember at Fintrac, staff around the world convene virtually to reflect on our knowledge - how we produce it and use it - for stronger agricultural development results. This is the second in a series about monitoring, evaluation, and learning as part of that effort (read the first and last ones here and here).

As database manager for the Feed the Future Ethiopia Value Chain Activity (FTFE-VCA), I am responsible for maintaining our monitoring and evaluation (M&E) database for routine monitoring activities about smallholder farmer acquisition of knowledge and skills. This means I am at the center of where field data comes together. From that data, we generate knowledge to facilitate data-driven discussions and decision-making for FTFE-VCA. As part of this process, I focus on ensuring the data remains at the highest quality. 

Analysis to Data Visualizations

Data is collected in real-time from across our technical activities, such as technical assistance, farmer field days and trainings, and market linkages. It is then reviewed and uploaded each day by a team of six data entry clerks across FTFE-VCA’s four regions – Amhara, Oromia, SNNPR, and Tigray. I perform a final quality control check on the synchronized data, including a multi-level manual review of source documents. I can then use SQL (Structured Query Language) to extract relevant data from CIRIS, prepare datasets for analysis, and finally, and most importantly, create sample visualizations in Tableau.

When our interventions were starting to ramp up in early 2018, we began putting our findings into data visualization packets with the goal of pushing the project team to get more comfortable with interpreting data. It is an important step before team members can adaptively manage using the information – they first need to understand it and trust it. Since then, I have been developing monthly visualization packets (focused on program targets across six core value chains - maize, coffee, chickpea, dairy, meat and live animals, and poultry). Visualizing the data helps communicate the findings to team members across Ethiopia and results in evidence-based decision making.

Discussing the Visualizations Leads to Actionable Decisions

For example, in early 2019, SNNPR was achieving only 50 percent of its cumulative targets and had just six months remaining to achieve them. We used our data visualizations to communicate this issue in the regional meeting with managers, M&E specialists, and agronomists. It set off decisive action with the regional manager assigning each value chain specialist to devise new strategies. They coordinated and collectively transformed the strategies into a regional plan for leadership approval. Because the new plan indicated a need for additional surge support, the field teams had the resources and vote of confidence they needed to achieve their outstanding targets.

To be able to facilitate such decisive action, I always try to make the visualizations as simple and easy to understand as possible because the audience is a mix of technical and non-technical teams. Tableau offers countless types of charts and graphs to best represent the data, and I select the most appropriate visualization type based on the analysis type (trend, pattern, etc.), and importantly, the comfort level of the technical teams. I export each into a PowerPoint slide, and focus on addressing one question per slide. If a slide depicts a trend, then I often share a more detailed analysis (or disaggregation) along with a concise take-away message. 

Actionable Decisions Facilitate Project-Wide Learning

The purpose of the learning component of MEL is to use the knowledge that we generate from the data. Applying the knowledge means that, as the database manager, I am a part of the project’s decision-making processes. The data doesn’t collect dust on a shelf but rather loops back into how we are performing, and it makes my job more meaningful and satisfying. Before the data visualization initiative, I felt that we were not always using MEL data to its full potential.  Now, our data visualizations drive technical discussions that translate into impactful programmatic changes. They’ve also created opportunities for collaborations between those of us who generate knowledge from data, and those who need to apply it to run the project more efficiently.

Although it might be difficult to measure the effect/attribution of using visualizations, it is clear to me that this initiative has played a key role in improving the project’s FY2019 performance. For one, it has created healthy competition among regions and across value chain teams. Additionally, regional managers and value chain advisors now use the data visualizations to plan new activities and motivate their field staff when targeting key beneficiary groups – which you will hear more about in our final blog of this series!