For about six months we've been running a survey on spending, income and food consumption. Thirty-two farmers around Bungoma, Kenya are using daily logbooks to record food spending and consumption in the household; a survey agent visits weekly to conduct an interview about non-food income and expenditure, and to swap out the old logbook for a new one.

We now have enough data to start asking interesting questions. I'm not a scientist and our numbers aren't statistically significant, but I'm comfortable asserting that these numbers are representative of the average family in our program. All averages of the data from the thirty-two families in the survey.

• Household size: 6
• Daily food spending: $1.10
• Weekly income: $18.38 (biggest component: general handyman work, $5.86 a week)
• Weekly non-food spending: $9.08

In terms of cash, the average member of these households is living on about forty-two cents a day. However, while our farmers have a surprisingly diverse set of income streams, their single biggest source of "income" is from the food they produce themselves. We've been running a food price survey in parallel with the expenditure log, so we can use the farmers' self-reported data on amount of food consumed from personal food stocks to calculate in-kind income (income from food sold shows up in the "weekly income" number above.)

• Weekly in-kind income: $14.46
• Weekly income, cash and in-kind: $32.84

So including in-kind income, people are living on about $.78 a day.

Some caveats: we expect large seasonal effects in income and expenditure, so these numbers will change as we get more data. (Big deal: we don't yet have income data from sale of farmers' maize harvests.) The first three months of the survey included only five farmers as we refined the survey format and the interview process.

Next up: some ugly Excel charts.