After months of great work by the partnership teams at Akvo and Water For People, FLOW settled into its new home with Akvo in March. Since then, we’ve been heads down, working hard to scale the system and meet the demand for new users. We reached one big milestone at the end of June when we released the Akvo FLOW source code on Github. Now it’s time to surface and tell you a little more about what we are working on.

Enumerators taking practice surveys at a waterpoint during a training in the Atebubu District of Ghana. Photo credit: Caetie Ofiesh

Akvo FLOW is a monitoring tool, and one of the most important aspects of monitoring a project is being able to see how the data changes over time. That’s how organisations know whether their work is effective or not.

The ability to track the performance of water points over time was one of the driving forces behind FLOW from the very beginning, so we’ve been steadily working on features that will achieve this. It’s straightforward to map and survey a point once, so the core question is: If you return to a previously mapped point, how can you update and add to the information associated with that point in the database?

Here’s what we’ve built so far that can help users update points on follow-up visits and build a historical record of project performance:

  • A Nearby Points function on the FLOW Field Survey app that allows device users in the field to summon a list of existing points from the database ordered from nearest to farthest based on their current location. They can then select an existing point if they see one that matches the one they are currently surveying, and bring that unique code into their current survey. These two surveys end up with the same unique code in the database so they can be associated with the same point. We’ve built something similar on the web versions of FLOW surveys, where users can choose to select a location for a new point from a map, or choose from a list of existing points.
  • Because a unique code is so important for point updates, we built a barcode scanning question type in FLOW surveys that calls an external scanning app (such as ZXing or Quickmark) on the device during a survey and records the barcode number. If organisations or communities want to install a barcode system for their projects, it immediately becomes much easier to have a unique identifier for points that can be used for updating. This feature has been on field trial with IRC in Burkina Faso, where they have been putting barcodes on pumps and giving people temporary IDs to record who is using the water points and when.
  • Support for a distance-based criteria for when two points should be considered the same point based on location. For instance, a FLOW user could specify that any two points whose GPS coordinates are within 5 meters of one another should be assigned the same unique code because it is highly likely they are the same point (there are some exceptions to this, especially in urban settings).
  • A Metric Manager on the Dashboard that will allow users to compare metrics (like tariff amount or user satisfaction level) from survey to survey and from year to year. Using metrics allows organisations to change the way they phrase their survey questions but still track the same pieces of information over time.
  • A unique code for points based on GPS location. GPS locations taken by FLOW field users are automatically turned into unique codes of letters and numbers. So (Latitude -1.65714566, Longitude 36.67799482) turns into: “244m1zqz.” We built it this way because most project points aren’t consistently coded, and using a location-based code ensures that they will always be unique (with some caveats [1]).
  • A Point Manager on the Dashboard where FLOW users can manually edit data for points that have been surveyed, including the unique code if they are the same point.

And here’s what we still need to do:

  • Build a good user interface for editing the unique codes, setting distance-based crtieria and manually reviewing points that are close together and may need to be separated or joined.
  • Design reports and visual displays of data for points that have been surveyed multiple times.
  • Field test with users to tweak our system and improve usability.

The usability and testing phases — bringing it all together — can actually be the most difficult to get right. So although we’ve made great progress, we’ve still got a significant amount of work to do, and we are working with partners to bring these ideas closer to reality.


[1] The resolution and accuracy of the GPS on mobile devices limits our ability to make sure every point recorded is accurate and unique. So in some situations, taking a GPS location on two visits to the same point may result in two different unique codes, even if the user is standing in the same location. In other cases, two points that are very close to each other but not the same could be given the same unique code. We will build review screens so that users can resolve these close calls during manual review.

Caetie Ofiesh is the product manager for Akvo FLOW.