Liveblogging the Zambian Elections

Phone Operators at the Casino/Data Center

I'll be liveblogging the Zambian elections from the Civil Society Elections Coalition (CSEC)'s base at the Taj Pamodzi from my own perspective. If you want to read chronologically, start at the bottom - I'll insert new content up top as I go. I apologize for any typos, etc - it's gonna be a long day.

2:05 AM

Signing off for the night. Ballots are being counted across Zambia; the process is wrapping up in many places but still has hours to go in others. Our doughty data clerks are taking the calls coming in. For the last while I've been staked out at our front entrance. It's critical to keep these spaces secure; everyone in here must have a badge, and visitors must be escorted. Too much important information to let random people wander about.

In a late-night triumph I found a "Bitter Lemon" soda, which is a refreshing change from the endless cups of chicory Nescafe. I've only had these in Africa; they're a bit like a tonic water with a shot of lemon concentrate. Probably don't need any coffee at this point. #twitch Now I'm going to head home and crash for a few hours; my colleagues tag in until 8 AM, and so on.

Thanks for following, folks. If you have questions or comments, don't hesitate to hit me on Twitter as @cdoten. Tomorrow will be a big day for Zambia - keep your fingers crossed that all goes well.

11:35 PM

Pulled myself another cup of coffee. Things are quiet around here now; we just finished one more round of calls to the supervisors to get a bit more information on how their polling center monitors did.

Counting of results is complete in a lot of polling centers now. The ballot boxes will then travel up to the constituency center (like a congressional district in size) and then will be summed up. The only call we're really expecting to come in is the number from those constiuency collation centers - 150 of them. They're all calling the same number. If you do the math with 5 minutes a call, it'll take longer than most folks feel like being on hold.

We handle it in a somewhat crude but effective fashion. When constiuency supervisors call, the operator on the Very Important Phone (VIP) will take their number, thank them, and hang up. They then pass the sheet of to another phone which will call them back. This both takes the load off the main phone and prevents them from needing to use up many valuable cell minutes.

I'm going to stick around here until 2:00 and then the next shift of managers will come through. From there we've got a solid six hours before we get back and the massive work of data-entering the 8000 general observer forms begins in earnest.

8:05 PM

Cleanup time. We're doing some initial analysis on the significant issues that our monitors reported through the course of the day. We're on our third shift; these unfortunates will be here manning the phones and computers as the first results are calculated. We're expecting them to stay all the way through until the next group arrives in the early AM.

There was another round of chicken and chips for dinner. I think that 90% of my caloric intake over here has been chicken, fries or donuts. Even during an election the work doesn't stop for supporting our other NDItech programs, so I'm going to take the next couple hours to catch up on that.

It's interesting and a bit sad to see the number of articles that immediately dropped painting a grim picture of sweeping violence. We'll see what the facts show, but I think there's a lot of people primed to jump directly to articles about the violence and anarchy of Africa.

6:01 PM

Polls closed!

Except for the people still standing in line, who get to vote. Or all the polling places where they started late and added more time. Of course, it doesn't really matter for CSEC's team, as the monitors will be there until the bitter end of vote counting, and we'll be back here at the data center gathering the information. But still, woo!

5:10 PM

One hour to polls close! I've been wandering about in my free time trying to take pictures and update CSEC's Facebook page. A really robust social media strategy would involve a lot more participation from regional leadership, monitors on the ground, and even clerks her in the center. Of course, given the level of internet penetration, it would be hard to justify, which is why they have not.

4:15 PM

The plural of anecdotes is not data.

Twitter is a wondrous creation. Never has the mind of man created a faster, more efficient, more interconnected method for distributing pure bunk. Rumors are flying fast and furious here in Zambia: polling stations ransacked, ballot boxes stolen, slips preprinted, election commission trucks driving into lakes. You can get a taste in the tweet stream for some common hash tags. Perhaps fortunately very few folks are on Twitter here; Facebook is far more popular, but even that is a small sliver of the population. Amid the chaos and tension of an election there are few facts for the first few days - the period in which restive groups are deciding to turn towards violence.

This is one of the things monitoring can help with; patterns are more important than individual events. An election commissions truck put to the torch is bad, but won't affect the election significantly. If that's happening in every single corner of the country, well, then we have a problem. A network of monitors across the country can verify the scope of the problem and reassure people that what their grandmother's cousin's neighbor in Lulapula reported isn't recurring everywhere. Or, if we are, we can provide a more solid base for what's going on.

As an aside, this is one of the risks of crowdsourcing. Even if you take care of the other methodological challenges, you still end up with a system where people are only motivated to report problems. If one looks at a map and sees nothing but pins indicating bad stuff that went down, you'll have a rather jaundiced view of an election that might have been of really excellent quality.

Speaking of crowdsourcing, bantuwatch.org is back up and accessible again from most Zambian networks. No idea what was going on.

3:10 PM

Time for Shift Two- the first group pulled their eight hours, so we've got a new fresh crowd of UNZA students in. Things remain relatively quiet. Since we're only recording incidents here, the vast majority of the work begins late tonight and tomorrow morning, when the first shipments of paper checklists come in from all our observers.

Logistics are challenging for these sorts of events, since you can't simply assume that everyone has the ability to get here on their own. CSEC hired a bus that makes a few stops across Lusaka to collect the data clerks; even so, some had to get up even earlier to travel some miles to the pickup points.

Things are likely to be a bit slower until polls close in just under 3 hours; some of the folks who arrived at 5 are slipping off for a catnap. I'm gonna get caught up on some other work.

1:45 PM

And now for something completely different: lunch. Chicken, though not Nando's today. DC friends might be amused to learn that Nando's is a ubiquitous fast food chain out here. First food today.

1:10 PM

Net shenanigans afoot in Zambia.

It's been an exciting hour around here with something very funky was going on with the internet. First I was attempting to reach an Ushahidi platform at bantuwatch.org with no success - but folks in other countries could. After testing from both the hotel network and from a GSM USB modem I could confirm it was simply inaccessible to us.

Then other sites started dropping. Twitter, Facebook, CNN - unreachable. Some major local media sites remained up, particularly the Times of Zambia. I didn't have a big enough library of sites to determine if it was an outage (or "outage") for cross-border networks.

There's a real challenge trying to figure out what the problem is when you can't get to a site. Most of the time it's a PEBCAK (stupid user error, which I often fail on.) Or it might be your network card. Or your WiFi hot spot. Or your hotel's connection. Or one of many ISPs. Or, occassionally, a coordinated national effort across all providers. It's not clear what this was, but it was defintely upstream from me.

At the time of writing, bantuwatch.org is still inaccessible for me.

11:50 AM

2 scoops Nescafe + 2 scoops cocoa powder = very poor man's mocha. Marginally more palatable. #twitch

Halfway through the voting period.

11:20 AM

Some VIPs have begun to flow through the center. We've got the biggest local non-partisan group of observers in the country, but we're not the only game in town. There are a lot of groups that do what is called "International Monitoring;" to set up one of those, take a high-profile group of VIPs and send them to a smattering of sites arcoss the country. While they don't have the breadth or depth of knowledge that something like CSEC's operation would, they bring a big media presence that captures cameras. In the tree-falling-in-the-woods sort of way, the most perfect monitoring information is useless unless someone is listening. Such international monitors often like to get a taste of what a really large domestic observation operation is like, and so we expect a number of them to come to the casino headquarters.

In addition, there's the steering committee members - these folks are the leaders of the already-succesful NGOs that pulled together to form CSEC. While they will be taking their own personal observations around the country, they will also want to check in at the center and see what their collective efforts have helped to create.

10:00 AM

All humming along well here. The server hasn't had any glitches yet and is having no problems with the load - possibly because it's a quad-core Xeon server that sounds like a jet taking off when you turn it on.

There's been some human error to deal with. Things that seem to be the obviously correct way to process information are not, of course, actually obvious at all. We always end up doing a lot of training with our data clerks. They got introductions to concepts and the system in their first test, they all trained on actual data entry with the simulation today, and we began with a careful walk-through of data flow. There's still problems when it actually comes time to *do* it, of course. The floor managers are keeping an eagle eye out and fixing issues as they arise. Data quality is paramount in this process, so we try and make sure there are various checks on the data and that each individual has a discrete, manageable piece of the overall workflow. We bought a lot of in- and out-trays.

9:00 AM

One of the biggest challenges of election moinitoring is actually being able to get in there and monitor. If obserers not able to get in that can cover all subsequent sins; this early in the morning this is the most likely incident to be reported.

Anecdotally, staff at the Taj report that turnout appears to be good. There's no real Twitter #hashtag for the election, since there's not much Twitter usage. Facebook, however, is huge - CSEC has a page as well. 4.5 hours in a new coffee heater arrived (the old one must have burned out. Weird! *whistles, looks around awkwardly*) and I've finally gotten my first cup of coffee.

7:30 AM

A flickr slide show of pictures from our casino data center. I'll add to it throughout the day.

6:10 AM

Polls have been open for 15 minutes now, and the phones are just starting to get calls on "Critical Incidents." Critical incidents are events that jepordize the integrity of the vote at some particular polling station. A few of the categories of things we're looking for include violence, polls opening late, refusal to allow monitors in, and so on.

Whenever someone has such a problem, we don't want to wait for two days to hear about it - so they call us at the casino data center. If it rises to a significant level we can contact the election commission or police or other monitors and let them know to be on the lookout.

A lot of the job comes to workflow and data management to get the info from our 8K observers into the database. One set of people are manning the phones and copying the reports on paper; others then take the info and data enter it. If there are questions or problems with the form, the monitor gets a callback.

I know I joke about coffee a lot, but it just about killed us all here. The urn was boiling away but, as it turns out, has no auto-shutoff. So it just vigerously continued bubbling and steaming. When I wandered back to the area I smelled something burning - the very hot, very empty urn. After jumping the counter and unplugging it we're all OK. Except me, who has still not had his coffee.

4:35 AM

Good morning! I'm going to attempt another liveblog of Zambia's election today.

It's way before dawn here. We were up past midnight taking care of the 1K details that need to be taken care of. There's still a few I'm handling - printing "In" and "Out" signs for the boxes we'll use to move paper, tidying up the place setting out the phones, etc.

And finding coffee. We're supposed to have hot water for the chicory Nescafe (*shudder*) that they provide. Unlikely to survive long on the handful of candy corn I brought in from the US.

I've just instructed the guards to make sure that absolutely no one without a badge makes it in to the casino datacenter. It's important that we don't have strangers wandering through here, either stealing information or disrupting our process.

Now, about that liquid sleep replacement...