While I am writing a business analytics report for one of my assignments I thought to write a quick post.
Firstly, we have to keep in mind that there is an acceptable structure for writing a report. If you’re analysing data to help your department make informed decisions, following a simple report structure can help you. I found this SlideShare from Sherrie Lee helpful.
Keeping it structured
I try my best to keep a logical order in my analysis and report. This allows me to understand my analysis better.
Based on chart 1.1. we can hypothesise that car models 1 and 3 have a lower average quality score than models 2 and 4, therefore, they receive higher numbers of type 2 and 3 complaints. To prove that, an analysis of the average quality scores against the complaints is needed.
You can see from the sentence I emphasised above that following a logical structure can help me think “OK, what do I need to do next and why?”
The first posts of my data science journey points out that although we need technical skills to help us analyse data, one does not need to panic.
When I started this assignment I was overthinking it a bit. Then I decided to keep it simple and use Excel and SPSS as the two main tools for analysing my data.
To be honest, because I know how to use the tools it doesn’t mean that I don’t need to invest time to get the results that I want. My main objective is to provide visuals that either prove or disprove my hypothesis. We need to use the right data points and the right charts to make sure that we convey the message we want to in the report. That is time consuming and we need to think a lot. So keeping it logical and simple can save us a lot of time and confusion.
One of the main purposes of the analysis report is to make meaningful recommendations. Let’s say we found out that there are problems that we need to address. Unfortunately, the analysis concludes that we don’t have enough data to figure out causation of problems. We can then recommend what actions or data we need to collect to help us find the causality of the problems.
Causality is not easy to find in the real world, but we can find correlations. We can then develop hypothesis based on the observed correlations and test this hypothesis.
The point I am trying to make is that if we don’t feel confident about what recommendations to make after analysing a set of data it’s OK. We might need further evidence and that could be our recommendation.
I haven’t done this yet, so I hope I will get it right. I’ll probably tweet out how I do when I get the results of my assignment.
I hope this short post is helpful to new data scientist. Let me know what you think, how do you approach your analysis and reporting? Do you have any tips? Share your thoughts in a commend below or via a tweet.