Akvo FLOW in Kajiado, Kenya (Amit’s viewpoint)


After numerous delays and a forced landing in Mumbai, it took me two days to reach Nairobi from Delhi instead of seven hours. But I was quickly stirred into action by Luuk Diphoorn’s intensive work schedule. My stay in Kenya was divided into two phases. The first was to work on the logistics of the Akvo FLOW training for WASH Alliance partners and support Luuk’s Akvo RSR training. The second was to conduct the Akvo FLOW training.

Above: Luuk’s HUAWEI 3G modem & router. Photo credit: Luuk Diphoorn.

This training was hosted in Kajiado, in Rift Valley Province in Kenya. In a way it was an appropriate site. The name “Kajiado” is derived from the Maasai word “Orkejuado.” which means “The Long River.” It seemed like an auspicious place to bring technology to water, which the Akvo FLOW training was attempting to do.

However, it was also perhaps not the best site. First of all, the training hall had only two working power outlets, and we needed to run 20 laptops on power throughout the training (including the LCD projector). A complicated set of extension chords was brought into play, but a fear of sparks flying kept us on tenterhooks.

Secondly, there was no provision of wireless internet in the training hall. This was the biggest challenge of all. Without Wi Fi, we would be unable to check for the Wi Fi MAC addresses of the smartphones the trainees had with them. The Wi Fi MAC address is a unique identifier (i.e. a signature) of each phone, from which they can be tracked on the Akvo FLOW dashboard. Without them we wouldn’t be able to remotely install the WASH Alliance surveys on the phones.

Thankfully, Luuk had carried with him a clever device from Amsterdam — a HUAWEI 3G modem & router. By using a 3G Safaricom SIM, we had our own Wi Fi system running.

We’ve started trying to take nice quality photos, outdoors, of everyone involved in training sessions, which we then upload and caption accurately in Flickr. It’s a really useful way to remember who is who and forge a long term connection. The pictures become very searchable, too.

In total, we had 16 trainees from various organisations. We were also joined by Francis Warui from Upande. Francis is using Akvo FLOW to collect data on water points in various districts in Kenya and his presence was helpful. Francis not only shared practical tips, but also took a session on how he was using Akvo FLOW to collect and analyse data.

The training went on for two days (indoors), and one day of data collection in the field. Some of the challenges of day one were:

  • Phone diversity: The range was astounding, from a very basic HUAWEI IDEOS phones to the latest Samsung SIII. One of the participants had a LG P350 – a good phone, but without a folder structure, making it impossible to upload survey or download data from it offline. The SIII was much more advanced than the gingerbread 2.3 O.S we generally are used to in smart phones, so it took us time and effort to figure out its settings.
  • Computer diversity: From operating software (Windows XP, Windows Vista to Windows 7) to different browsers (Mozilla, Chrome, Internet Explorer) and different versions (for example some had IE 7, some had 8), to processors (some 32 bit and some 64 bit). As a result, we needed the latest version of Java to run reports and other functions on the dashboard. But given different specifications of systems, we had to run around fixing issues that kept cropping up.
  • Unreliable internet: This Akvo FLOW training took place entirely on 3G mobile networks. All participants had individual 3G USB modems, and most of them were on Safaricom. But for some reason, all mobile service providers, namely Airtel, Orange and Safaricom were very unreliable. This meant frequent loss of internet access, which slowed us down.
  • Lack of availability of smart phones: Though some organisations were supposed to purchase their smart phones before the training, there was a last minute dash, which meant three new phones arrived only by lunchtime on day one. Setting up those phones and letting users warm up to them took some time as well.
  • Creating Gmail Accounts: This also took considerable time, especially given slow or no internet.

We have the following to take away from this as a pre-training checklist:

  • Cheap smartphones work beautifully as far as Akvo FLOW is concerned, and their increasing penetration is a great opportunity by default.
  • Make sure all phones are in place before the training, familiarise people on locating Wi FI MAC addresses using instructional videos.
  • Before training starts, get specs of peoples laptops, including OS, processor and browser version, if possible make sure all of them have upgraded their systems to the latest Java.
  • Make sure dashboard users have their gmail accounts online and ready.
  • Make sure Akvo FLOW training takes place in a venue where there is a reliable broadband Wi Fi and some electricity back up.

In the previous Akvo FLOW training of BSP- Nepal, we managed to do most of this, largely because the training was for one organisation only. Since this training was of consortium partners, it was difficult to get all this in place before the training.

As the training on managing dashboards came to an end, trainee’s created their own questions on the dashboard and uploaded it to their phones. However, some phones did not get the surveys installed. Francis suggested that we try the “reload all surveys” method, which forces the phone to talk to the dashboard, and it worked like magic.

At the end of day two, as we were finalising the survey for field data collection the next day, a new problem emerged. The questionnaire that was to be used was largely for partners and needed substantial tweaking to convert it into a WASH group related survey. Some midnight oil had to be burned and we got the approved survey edited and ready for installation.

Day three was perhaps the best. The data collection went off smoothly and we captured Joseph from NOSIM sharing his experience with us, along with a few other trainees in the act of collecting data using Akvo FLOW. After field-testing, we returned and took a look at the data. As always, that’s where the “automagical” aspect of Akvo FLOW comes to light, and all the sweat, frustration and exhaustion was justified when trainees saw how their data, submitted just an hour earlier, was available to them analysed on Excel sheets.

As we concluded the training, the support and interest of the participants was inspiring, to say the least. Later, Luuk and I sat down on the porch of our guesthouse with a Tusker Malt in our hands, watching a lilac sunset unfolding across the savannahs. We had done Akvo FLOW training at the frontier of what’s practical. And we had survived.

Cheers to that.

Amitangshu Acharya is a consultant, Asian programmes, for Akvo

Related: Akvo FLOW in Kajiado, Kenya (Luuk’s viewpoint)