I seek some guidance regarding which method(s) are best suited for evaluating a dataset I'm working on. My data consists of newly launched pharmaceuticals (~100 products), with sales data for these between Q1 2015- Q3 2018 (4 meassurements per year, so 15 meassurements). Some of these have data for the whole interval, some only started selling later (for example in Q2 2017).

Together with this, I have ~10 different events (Event1 - Event10) that can have happened at any time, sometimes several times, for the product in the time interval 2015-2018. I am basically trying to figure out if an event causes a change in sales, and in which case what event affects the sales the most.

Right now I have An Excel-sheet with rows of product names, lets say of length m, with sales data for Q1 2015 to Q3 2018, basically an m x 15 matrix (Example in image below).

Example of how the data looks like

The event data are for each product in 2 columns, 1 for the event type, and one for the date of the day the event occured, basically m number of these matrices, with 2 column but of different lengths (depending on how many events occured for that specific product) An example of how the data looks like is in the image below.

Example of how the event data might look like

My question, I think, comes in two parts:

First, both the sales data and event data are as a function of time, only that they don't take the same time step. I assume that this has to be altered for the datasets to be evaluated against eachother. Any tips on a good way of doing this?

Secondly, what methods should I look into for evaluating if an event at one point in time causes a change in sales going forth? Basically, for each product a combination of events 1-10 can have occured at some time during the time interval, and I want to figure out if a certain event somehow affects sales for that same time interval.

  • $\begingroup$ I would start by plotting the series, maybe with linear interpolation between observations. Then mark the time of the events on the graph $\endgroup$ – kjetil b halvorsen Mar 9 at 18:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.