Data mysteries

Data processing, it’s mind numbing sometimes. I guess it depends on the dataset you’re working on, but in this case it’s just a formatting thing. Literally copy, paste, rearrange, things like that. The issue is the volume of data you’re working with. The whole thing can take hours just to get done properly, but with any luck you’ll only have to do it the once. Most of the time I’m not that lucky, the last dataset I worked with for my Co-PI I reformatted five or six different times. While that isn’t the topic of the day, it’s a good way to start the conversation of the day.

I genuinely enjoy my research. On occasion my Co-PI and the lab director will laugh at my excitement when we’re doing something cool. It’s all in good fun and my Co-PI gets just as excited as I do. That is why when I had more on my plate than I knew what to do with, I couldn’t say no to doing the analysis of the dataset we collected from a rather interesting experiment. That would be the dataset from the “When you don’t want to say no” post. Unfortunately it got put on the back burner because there wasn’t enough hours in the day to do everything I needed to get done and hit all the deadlines I needed to hit. Well I hit every single one of them, sometimes painfully. So now it’s time to crack open this dataset and see what we get.

I’ve learned a lot over the past two years of working with my Co-PI. He’s incredibly hands on when he’s teaching, but gives me the freedom to do the analysis or think the way I want. Those of us in the lab tend to think of it as working WITH him, not FOR him and we’re all better for it. The lab dynamic is great and it makes me excited to know that I can propose something and have a serious discussion without getting blown off. That isn’t to say my main-PI is a bad guy, he’s just so incredibly busy that he literally doesn’t have time to do the same stuff with the lab generally speaking.

My Co-PI suggested it would take ~6 to 12 hours to look over the data. I knew he was being incredibly optimistic, but once I sat down to work it took almost that long just to get the data into a format I could work with and aggerate the subjects we had. The trick is more data is still being collected, but it never hurts to peek behind the curtain to see what’s going on in the interim.

As usual I can’t go into details of the dataset, but that doesn’t mean I can’t share my excitement about it and share how incredibly painful the dataset is to get into a format that’s easy to play with. I can also say to a point that we’re trying to map neural circuitry based on recorded responses, but I can’t give details beyond that. The trick is everything is complex, the brain is complex, the spinal cord is complex, there’s pathways for reflex’s, volitional movement pathways, things that interact with one another and override each other (to a point) so it’s tricky to isolate any one thing and the interactions between circuits and what not needs to be taken into account.

I’ve been tasked not only to clean up the dataset and get it into the right format, but also to make these inferences based on what we see. My Co-PI gave me a short list of interesting questions he wanted to answer with the dataset, but there’s no direction and I’m not limited to his questions, it’s up to me and what I want to show. Well, I’m happy to say after yesterday I got everything in the proper order and I’m ready to start asking the dataset some questions.

That’s just a fancy way of saying I’m ready to plot different conditions. My process is to formulate a question or a hypothesis that I would like to answer using the data and then plot the information from the dataset I think will best let me answer that question, so in the literal sense I’m asking a question and the dataset is providing the answer. Once I find something “cool” I can expand on it or go a different direction depending on what I come up with.

My Co-PI wants me to make a very interesting presentation for him using this data, or at least make it look good. We have a “rival” group that makes incredibly intricate visualization for data and we joke about making ours look as good as theirs. Currently I told him we would discuss it today, but I’m slightly behind schedule and only just started interrogating the data, so while I’ll have answers, I won’t have anything pretty. Once we discuss it further though I can get into the data visualization portion of the process and make something unique that looks good. I have a few ideas that I plan to run by my Co-PI later today.

This is the part I really love, getting to figure something out. It makes all the boring stuff worth it. I don’t quite know what we’ll find here, but I’m excited to get the chance to find out!