The activities around processing big data involves huge number of calculations and research. This is purely science project in comparison to the regular development cycle. In the initial phase there are only guesses which must be checked and proved. All guesses might not work in the end. Nevertheless, if you get some good results you should be able to present them to the audience which does not have deep understanding about statistics and technologies you used during the investigation.
If you’re presenting your project to people who is interesting in its results you might not need to show all details of your work. Usual topics for data science project might look like the following:
- Business objectives
- Data source
These topics are good for somebody who does not know about your subject and goals, for example, somebody outside of your company. In opposite case they are too abstract and have redundant information. So let’s make the information concise and give only necessary information. What others might be interested if they know the subject are the following:
Thus you’ll show the results, how they were get and how they were validated. That’s it.
I would recommend to give only final values and brief clarification of the methodology. For example, you need to show and talk about validation of your models, why you chosen this model from others. The correlation between experimental results and real data can prove your choice and show that you’re on a good track. You might also need to explain the validation approach. Is it just real-vs-estimated data or cross validation, or something else. The explanation of statistics is not needed for the audience.