Maybe you’ve long held aspirations to work in an analytical environment—you enjoyed logic, math made sense, and you felt relatively at home with the idea of probabilities. Or perhaps you found interest as you did your research and felt intrigued by the work completed by statisticians or data scientists. It could also be that while you didn’t enjoy math while growing up, you matured your way into it and you’re quite comfortable with it now.
Congratulations, you conquered what might be one of the top necessities for success as a statistician/data scientist: not fearing math. But there are many other skills you need to possess or develop beyond the ability to conduct different calculations that will help you solve unique problems. Here are the top ten basic statistics concepts that any statistician/data scientist should know:
1. Statistical Computing
While it’s possible for statisticians or data scientists to team with developers, any statistician/ data scientist who can program (R, Python, etc.) has a definite advantage in terms of flexibility and innovation in the design of an analytical technique. Familiarity with statistical software (such as SPSS, SAS, Minitab, etc.), used by various companies and organizations could give you an edge.
2. Probability
The foundation for much of statistics is rooted in probability:
- The chance of an event occurring
- The conditions that must be present for an event to occur
- What influences the occurrence
- To what degree can we predict any event
A solid understanding of probability is key to determining, for example, if a result you found in the analysis is likely due to chance.
3. Problem Solving and Design
A top statistician/data scientist will not only know how to build a statistical model, but they will also understand why that particular model is the most effective for solving the problem at hand. They will be able to hear the problem, synthesize their understanding of the context and the needs of the organization, and design the most accurate and appropriate solution.
4. Sampling
The sample used to answer a business or research question is at the heart of any analytical design. The type of technique used to acquire the sample drives the type of research design, which drives the appropriate analytics to employ. For any statistical procedure to work, it’s necessary to have the right sample size and know how to find it. Knowledge of sampling and the inherent issues present within sample types yields the most valid and reliable results.
Learn More About Our Statistics Program
5. Descriptive Statistics
Descriptive statistics enables us to summarize our data set for better understanding and interpretation. It also assists us in determining the type of distribution the data set forms. Central tendency statistics show us the center of our distribution:
- The mean (the sum of the data set divided by the number of observations in the data set)
- The median (what number appears at the middle of the data set)
- The mode (the most frequently occurring response)
Variance indicates the spread of our data—the standard deviation, for example, shows us how far the responses are from the mean in a data set, if they mirror our target population and to what degree might we be able to generalize the results we create from this sample.
6. Predictive Modeling
Predictive techniques such as regression produce critical statistics for data science. Regression models can take marketing questions (who is most apt to buy this designer handbag?), or clinical concerns (who is most apt to have a heart attack?), assign weights to meaningful predictors, and permit for accurate predictions as to spending habits or at-risk clients.
7. Group Differences
Choosing the right statistical test may be a challenge that requires the right test selection process to compare two or more groups for statistical differences. Statistics for group testing include such techniques as a t-test or an analysis of variance (ANOVA). This type of statistical test can also assist in discerning what variables might interact with each other to produce an effect (would the results be different if the participate was male or female?)
8. Classification
This data mining technique enables the statistician/data scientist to reduce large amounts of data into meaningful dimensions or factors by way of principal component or cluster analysis. Decision trees—models of decisions and possible outcomes—fall under this category as well.
9. Hypothesis Testing
The ability to create a testable hypothesis is central to discerning if there is a statistically significant finding to answer a question.
10. Communication
Communication is a basic statistical concept. You’ve used your skills and capability to design a study and answer a pressing research or business question—the most important skill to utilize now is your ability to communicate and share your findings. Are your results clear? Do they impart the necessary information for decision-making? Your results could be accurate, meaningful, and robust—but useless if not well communicated.
The ability to summarize, generalize, infer, and communicate are key components in a statistician/data scientist's ability to be successful. If you are ready to learn statistics or integrate analytics into your organization, one place to start would be pursuing an online masters degree in applied statistics. To learn more about Michigan Technological University’s online Master of Science in Applied Statistics program, request information.
Learn More About Our Program
Sources:
- https://towardsdatascience.com/10-things-every-aspiring-data-scientist-needs-to-know-3db18cce8eb4
- https://towardsdatascience.com/the-5-basic-statistics-concepts-data-scientists-need-to-know-2c96740377ae
- https://medium.com/cracking-the-data-science-interview/the-10-statistical-techniques-data-scientists-need-to-master-1ef6dbd531f7
- https://stattrak.amstat.org/2012/06/01/successfulstatistician/
- https://www.forbes.com/sites/quora/2017/06/15/what-are-the-top-five-skills-data-scientists-need/#b3f9487c0c0e
- https://www.cio.com/article/3263790/the-essential-skills-and-traits-of-an-expert-data-scientist.html