I am currently teaching a class on Computer Applications in Behavioural Sciences and I would like to know the principles of data analysis.
From my search, I only came across the three by Benjamin Miller (2001). Beyond statistics: A practical guide to data analysis. (P1).
i.e.
a. Frame questions in terms data rather than in terms of software.
b. Look at data before computing anything (be alert for errors, outliers, skewed distributions etc).
c. Don't do categorical things with quantitative variables or qualitative things with categorical variables.
Any other principles?
Dear Anthony Muriithi Ireri : there is no single answer but I would tell you one very important principle for data analysis, data cleaning! Spent majority of your time in cleaning your data before starting the analysis for effective and efficient analysis.
Just add "big" to data and you get the 5 attributes requested for big data
-variety
-velocity
-veracity
-volume
-value.
You will find many industry white papers on big data.
However, statistics preexists big data. I recommend you look at Statistical Disclosure Control (topic of research and white papers).
Data quality also appears in Artificial Intelligence as prominent.
The best example of historic quality data is the time series of temperatures measured and recorded in Padua (Padova, Italy) for 450+ years.
The quality of measurements in physical sciences is paralleled by criteria for tests where humans are observed. There is literature available on this.
Hope this helps
Dear Anthony Muriithi Ireri : there is no single answer but I would tell you one very important principle for data analysis, data cleaning! Spent majority of your time in cleaning your data before starting the analysis for effective and efficient analysis.
- Badges
- Science topic
- Similar topics
- Computer Science
- Computing