Recollections from before the rains…Data Ache in the Borderlands

It’s still red-dusty, but finally cool when I open my eyes. The horns are beeping in the street outside. The town is getting back to business after the holiday. The Paramount Chief’s funeral stopped all work on friday. But the fan has stopped. The generator must have run out of fuel – probably around dawn, I suppose.

What a comfortable bed. We did not holiday yesterday. In fact we worked the whole weekend. And one of the products is this room I am staying in. I carefully pull the bednet out from under the mattress so as not to tear it. A blue decorative net. There are burns and stain on the simple ‘sepia’ african furniture. Everything I look at is monochrome. Even the bars welded onto the window (ex-concrete reinforcement bars) are painted in this cream. And the red dust of the forest adds accents, despite the obvious scrubbing the three sons have been doing since our recce on Friday. I was going to move into this ‘proper african hostelry’, as I have called it, on Saturday. We had found it through the grapevine, and visited to cut a deal. We were led into the big open downstairs cooking area – pots on charcoal on the floor, stools and cauldrons around in the mud floor of the courtyard. I had insisted on seeing the generator, to judge the veracity of ‘24 hour electricity’. Then upstairs a series of tiles concrete corridors led into rooms and balconied common spaces, with shutters against the west african heat. Echoing tiled floors. The bed I am still lying on is enough to sleep a family on. Massive. Shabby. Clean Comfortable. Like the hostel. They had to close this place during Ebola. We are their first significant customers since.

I’m tired from yesterday’s long ride into the bush. Last saturday, Ivan had been monitoring our mapping surveys as they arrived at the server from the pillion seats of motorcycles surveying the borderlands. The Africell network sends the surveys from the african bush direct to some  cupboard in a back room in a New York high-rise. Data. And now, in Dakkar, Senegal, it wasn’t making sense.

‘Gentlemen, we have a problem.’ came the WhatsApp message, just as I settled down to a longed-for cold beer at the end of a relentless week. One of our mappers was submitting bad data. Ivan had clocked it in one village. Wrong name, wrong area. But if this was the case, this surveyor could have been getting it all wrong for weeks. It is hard enough to get to these villages once, let alone for one of us to go around double-checking more than the ‘ten percent’ quality control checking we do. But without quality, our project is useless. Just another incomplete mapping attempt. This is what makes us different from ‘top-down’ mapping projects. We are comprehensive and exhaustive. So we can get around the problem that lack of complete data can mean all the data is useless in a set. People are actually inputting their own neighbourhood data, from the field, via their smartphones. The surveyors are from these very neighbourhoods. Our data is meant to be the best.

But here was the problem: somehow, the wrong names were being attributed to villages. A cross-reference to another surveyor had been made. The only way to way to deal with it properly was to find out which surveyor was correct by doing a survey myself, on-site. So yesterday, during holiday, I had got on a motorbike out into the border sections where this problem had occurred. It was a rough day, and we lost a surveyor whilst we consider re-training him.


Photo: Rupert Allan
Lifting Motorcycles through Blocked Roads

It was a tough day. The roads make biking exhausting. Especially pillion. For five hours on a 125cc motorbike down bush tracks. There is seldom a helmet, and the roads are spine-compressing. In the afternoon, I was feeling nauseous, and worried that I had what Randy had last week – Malaria. The week had included extracting him from the isolated hill town of Baama, whilst keeping the mapper-training going. So a painkiller and some doxycycline and a big bowl of groundnut soup with meat had just made me start feeling better when the ‘data-ache’ news came through.

As I start to wake up properly, I remember today’s new issue starting to rear its head: the surveyors in the north are not uploading any data at all! Why not? Three days without uploading. The money we have for this is from Ivan’s savings. If we don’t get data, and it is lost or seriously wrong, the project will fold. It is critical that we get it right for this pilot, so that we can try to get financial support for the next stage.

And this is what I will deal with today. Wish me luck. But first, I must pay the Mama of the hostel. And get out into that insistent-sounding Sierra Leonean street…


A truck sheds its load on the main North-South road, blocking the arterial route for all but motorcycles, when an over-ambitious HiLux tries to pass…

Photo: Rupert Allan