Lets face it, data analysis reports, whether youre writing them for universities or for big data, are intimidating. Theyre also not a great deal of fun to write. I asked some people where theyd rank writing one and it came in just above going to the dentist. Thats not a good place to find yourself (and here Im talking both about the list and the chair).
You know what the crazy thing is? Theyre actually not that hard to write! Like so many things in life you just need to know where to start. For that reason, I thought Id write you up a quick article so that the next time you at least know how to get it over with as quickly as possible. And then, maybe youll start to enjoy it more and it will become only as unpleasant as being woken by your neighbor drilling holes in the wall. We can only hope, right?
There is no one right way
The first thing youve got to realize is that there is not yet one way to present your data. Admittedly, thats unfortunate. It would probably be helpful for everybody if there was some standard way to do these things. Thats not the case, however. Nor will it happen soon, as different companies, sectors and data presenters want to do entirely different things.
What that means is that there is no official wrong way. These are of course the unofficial one, which is where you hand over your 10,000-word report, people stare at it for a while, point to the first sentence and ask what does that mean?
Follow by example
The best thing to do is to find data analyses reports that you yourself find clear and easy to understand and to not just read them, but keep them on file. Then, when youre writing up a report, pull out one that is close to what youre trying to do and follow their example as closely as possible. This works surprisingly well, particularly if youre not that confident what youre doing is the right way to do things.
Whats more, if you do this a number of times without doing anything else, youll still start noticing underlying patterns and ideas that you can use.
Dont be afraid to ask
There are tons of very smart people online at such places as Quora and Linkedin Groups who are willing to answer even the most complicated questions, provided youve demonstrated youve at least tried to understand the concepts.
So no asking how do I write a data report? as then people will call you lazy. If on the other hand, you want to know if you should use a regression or a Linear Model for a specific analysis, then theyll gladly answer that for you.
Pay attention to the right criticism
Note that quite often youre going to get criticized. Thats just the name of the game. There are always people out there with opinions and the desire to share them. Make sure you pay attention to the right ones.
If the person delivering the criticism has no understanding of data or how to present it (and lets face it, there are a lot of those) then you can generally simply say thank you for them being so kind as to give their opinions and move on. Of course, if the person that is doing the criticizing is paying for the report, thats a different story. Unfortunately, then youre going to have to accommodate their criticism.
Know your content
When youre going to write a data report, make sure that you know the content. With that, I mean that youve actually been the one to do the analysis. It should always be the person who has done the analyses that write the content. Only in that way can you be sure that the data presentation isnt distorted.
Start out by describing the data. Who was the population you analyzed? How did you collect the data? What were the conditions, if any, for you excluding samples from your data set? Did excluding them make a substantial difference to your results?
At the same time, you dont want to go into too much detail here. Include whats relevant (how results changed as a result of different treatments) and ruthlessly cut down on whats not (what program you used).
Know what matters
A data report, just like any other report, has a message. It is trying to convey an idea. Of course, this is often more difficult when youre writing a data-driven report as data rarely cares about what you want to say and can therefore not fit your story exactly.
Nonetheless, you should focus on telling your stories (while not deleting the data-driven caveats). The natural structure of abstract, introduction, method, results, conclusion, thats generally used for these kinds of texts will help in that regard as it will steer your story.
The abstract is vital
A lot of students will write the abstract right at the end, when everything else is done and when theyre thinking about their friends down at the pub. Dont be like those students. The abstract is probably the most important part of your entire paper. Its certainly the part that is going to get read most.
The abstract is not a wow, pazaaz! piece. Youre not trying to impress your audience with it. Instead, youre trying to summarize the main findings of your data and the conclusions you draw from there. A well-written abstract will make the rest of the paper much easier to digest.
Remember to cater your abstract to your audience. If theyre looking to find out about a specific result, then include that! This way, what they want is right there at the beginning. Yes, that does make it less likely that theyll read the whole report, but then if you only have to explain the abstract, that makes life so much easier.
Numbers and words
Remember that not everybody understands data and what it means. For that reason, when you write out a result, dont just write it out in numbers, also write out the results in words. In fact, if you really want to do it correctly, you want to write it out twice the first time to translate the results into words pretty much one on one, the second time to explain what that actually means.
In this way, both the numerate and the number-illiterate will find it easy to understand what youre trying to say.
Last words
The most important thing to remember is that you can ask for help. If youre struggling with an analysis or how to write it out, go online and ask people! Youll be surprised how quickly people will get back to you if your question is specific enough.
Then, you can move forward confident in the knowledge that your report is as understandable as it can be. And thats a great feeling. And who knows, maybe once you stop being so intimidated by presenting the numbers, you can finally get that position as chief data officer!
