How I've Been Digging Into WVDOC Daily Incarceration Data
The West Virginia Department of Corrections posts daily custody counts on their website, and at first glance it looks like a boring spreadsheet. It isn't. Once you actually pull the data and cross-reference it month over month, you start seeing shifts that don't get discussed much in regular news cycles. I've spent the better part of two years pulling this data, mostly because I got tired of people making claims about incarceration trends without backing them up with actual numbers. The public dataset is there, but it's not organized in a way that makes analysis easy. That's the first thing you need to understand before you start.
West Virginias Rising Tide Uncovering The Patterns Behind Wvdoc Daily Incarcerations
"West Virginias Rising Tide" started as an investigative project that examined WVDOC data to show how daily incarceration figures have changed over time. What made it useful wasn't just the headline numbers, but the methodology behind how they tracked things like length of stay, admission reasons, and demographic breakdowns across different facilities. Here's the thing most people miss when they look at this data: the daily custody count is not the same as the committed population. The custody count includes people in regional jail facilities, work release, and pre-trial detention under DOC oversight. The committed population is only people serving sentences. These are two different denominators, and mixing them up will give you wrong conclusions every time. I learned that the hard way. In 2023 I ran a comparison using the daily custody figure and concluded that incarceration was rising faster than it actually was. Once I separated out the pre-trial and regional jail population from the committed offenders, the trend line flattened significantly. The growth was happening in the detention side, not the sentencing side.
Getting the Data
WVDOC publishes daily data through their Offender Information Search System and their public data portal. The main page you'll use is their daily custody report, which updates every business day. It's not automated, so there can be gaps on holidays or during system maintenance. I've found the most reliable approach is to download the CSV exports directly rather than trying to scrape the webpage, because the page structure has changed a few times and scrapers tend to break. The data fields you actually need are the facility name, date, total custody count, committed population, and the pre-trial or local detention segment if it's broken out. Sometimes the breakdown is there. Sometimes it's not. I've had to cross-reference with quarterly reports from the state's judicial branch to fill in missing segments during those periods.
Get the Full Details
What the Numbers Actually Show
Looking at the long-term trend, West Virginia's committed population has been relatively flat or slowly declining over the past decade. That's the bigger picture most people are looking for. But the daily custody count tells a different story because it captures flow, not just stock. When I started tracking this regularly around 2021, the daily counts showed noticeable spikes that corresponded with policy changes at the state level, particularly around probation and parole violations. One counter-intuitive thing I found: during periods when the committed population dropped, the average length of stay actually increased. That sounds contradictory until you think about it. If you're releasing people earlier but also holding more people on detainers or immigration holds, the headcount can stay steady or even rise while the underlying sentence population shrinks. The data doesn't always tell you why a number changed without looking at the administrative notes that accompany quarterly reports. Another thing that trips people up is the difference between peak custody and average custody. A single bad month with a facility at or near capacity will look like a trend if you only look at monthly snapshots. I switched to running rolling three-month averages and that smoothed out a lot of noise that was making the data look more volatile than it actually was.
Practical Problems I've Run Into
The worst issue I've encountered is when WVDOC consolidates facility reporting. A few times over the past couple years, they merged or reclassified certain regional jails in their daily reports without clear documentation. The facility codes stayed the same but the population attribution shifted. If you're tracking individual facilities over time, you'll see sudden jumps that aren't real. The workaround was to pull the annual performance reports from DOC and match them against the daily data to identify where reclassifications happened. Takes extra time but prevents you from publishing incorrect trend lines. Data completeness is another problem. Not every facility reports every day. Some facilities skip weekends. During the pandemic there were extended gaps that required interpolation if you needed continuous daily series. I just flag the gaps and don't fill them in. Interpolating prison data feels clean but it's misleading because incarceration doesn't change linearly.
Tools That Actually Help
I use Python with pandas for the heavy lifting. The basic workflow is: download the monthly CSV bundle, merge the files, handle the facility code reclassifications using the annual reports as a reference, calculate rolling averages, and then export the cleaned dataset. It takes me about 45 minutes to process a full year of data from raw download to clean spreadsheet. Setting up the reclassification mapping takes the most time, but after that the pipeline runs fast. If you're not comfortable with code, the data is simple enough that Excel can handle it for shorter time ranges. Just be careful with the merged cells in the original downloads. DOC's formatting is inconsistent between quarters and you'll waste time unmerging cells if you don't catch it early.
Limitations You Should Know About
This dataset doesn't tell you everything. It doesn't include county jail data that falls outside DOC jurisdiction, so your analysis of "incarceration in West Virginia" will always be incomplete if you're only looking at DOC numbers. The county jails hold a significant portion of the state's pre-trial population, and those numbers don't flow into the daily custody report. The data also doesn't capture releases that happen outside the reporting system, like transfers to federal custody or out-of-state placements. If you're trying to build a complete picture of population flow, you need to supplement with Bureau of Justice Statistics data and state court records. No single source has the full story. Finally, the daily custody count is a snapshot, not a rate. A rising number could mean more people are being incarcerated, or it could mean the same number of people are staying longer. Without length-of-stay data, which requires pulling individual offender records, you can't distinguish between those two scenarios from the daily count alone. That's probably the biggest blind spot in this dataset and it matters a lot when you're making policy claims.
