People interested in Pennsylvania’s Marcellus Shale data frequently ask me what they think is a relatively straightforward question: how many drilling permits are there for Marcellus Shale wells in Pennsylvania? As it happens, this is actually a somewhat complicated question, as there are numerous paths to finding the answer, all of which will lead to a different result. Consider, for example:

• If you go to the Well Permit Workload Report for the week ending 1/20/12, the answer is provided for you: 9,883. But…
• …if you search the Permits Issued Detail Report using the same end date, you will be given the result since the year 2000 as 9,868.
• The above item is particularly confusing when you realize that it consists of 11,209 rows of data. No problem, you say, I’ll just use my Excel skills to pivot the data by it’s unique API number, and as of today’s data we’re down to a mere 9,880 (plus two items with the wrong numerical formatting altogether).

So which number is right? In my opinion, none of them. If you follow the API link above, you will see that there are serveral components to the API number, which generally is in the format AB-CDE-FGHIJ-KL-MN, where each letter is represented by a digit ranging from 0 to 9. Here’s what they mean:

• “AB” represents the state code. In Pennsylvania, the code is “37”, but it is not included on the DEP dataset.
• “CDE” is the county code, which is alphabetical, and starts at 001. So for example, Allegheny County has a code of 003, since it is near the start of the alphabet, while Washington County is 125, which is near the end.
• “FGHIJ” is the unique well indicator. In theory, this allows for 100,000 wells per county to each have their own number. If you count the distinct combination of county codes and unique well indicators, there were 8,942 well permits as of 1-20.
• “KL” indicates the directional sidetrack code. This could represent multiple horizontal components of a well, so there is some wiggle room for argument if you want to consider each horizontal segment to be its own well. I argue against it, as the language talks about there being numerous horizontal components to a well, but for the record if you include it, then the number is 9,638.
• “MN” represents the event sequence code, which includes modifications to existing wells that also require permit actions. The number of distinct Marcellus Shale permits if you were to include the directional sidetrack and event sequence code would be 9,878 as of the 20th of January.

So…which number is right?  My interpretation of what the code means is that to count the number of wells in any given state, you should include all of the three digit county codes and all of the five digit unique well indicators (or “CDE-FGHIJ”, as described above.)  As of January 20th, that number was 8,942 for Marcellus Shale well permits in Pennsylvania, and as of today, that number is 9,005.

Sometime this year, I expect that the number of total Marcellus Shale permits in Pennsylvania to top the 10,000 mark.  But if that claim comes within the next week or two, my opinion is that it isn’t an accurate representation of the data–even if the claim comes from the DEP itself.

The following charts contain data through the end of January 2012.  The first is based on 11,297 permit actions (or records on the permit report), while the second is classified by unique 8 digit well API numbers.

by Josh Knauer, CEO of Rhiza

At Rhiza, we love to experiment with new ways of visualizing data that help tell better data stories. In most of our work environments, using data is kind of difficult and visualizing is usually left to data experts. We’d love to see a future where sharing data visualizations (maps, charts, explanations, etc) is as easy as recording and sharing a video on YouTube. Not everything produced will be stellar in quality, but at least we’ll all be a lot further down the road towards breaking down the traditional data silos and moving data aggregation and visualization solely out of the hands of database admins and graphic designers. We’ll still need those folks, their jobs will just get a lot more fun!

To this end, when I saw a data animation created by John Detwiler that showed the spread of drilled Marcellus shale gas wells in Bradford County, I wanted to create my own data animation telling the same story, but for the entire state of Pennsylvania… Read more»

One of the regular users, John Detwiler, recently took the mapping application that is part of the platform to an even greater level. He exported wells drilled data by year into Google Earth to create an animation of the shale gas wells drilled from 2007-11 in Bradford County, PA. Check it out: