Statistical Tests You Need to Know in Economics
Economics 101 for data scientists
What tin information scientists larn from economists?
- Accept you heard of the term econometrics before?
- Do you know that mathematical modeling and statistics are fundamental to quantitative economics?
- Practice you know that some of the biggest data repositories are maintained by economic inquiry organizations?
Just looking at the championship of this article you might exist wondering, really? What economics has to do with data science? A biased economist, I'm not here to tell the great things near economics, rather give an insider view of how economics works — as a way to demonstrate the connection with data science. I'll brand this connection in terms of (1) the problems solved, (2) the methods used and (3) the tools practical. Hopefully yous'll be able to draw some parallels with data science principles, tools and practices along the way.
Bug solved
Economists invented an imaginary homo — Homo economics or the "economic man" — to model human beliefs. That's all economists do, understanding human beliefs with economic theories and using some analytical tools and techniques — many of which are standard statistical and mathematical models.
Economists invented an imaginary man being — Homo economics or the "economical man" — to describe homo behavior.
Economic science bug are broadly brought under ii sub-disciplines. Microeconomics focuses on the interaction between basic edifice blocks of the guild and the outcomes of the interaction, while macroeconomics zooms out of the smaller details and sees the globe as a large interconnected organization.
Permit'due south say, a company is launching a new product. If they gear up the price too loftier, few people will buy it. On the other hand, if the cost is also small, a large number of consumers need to buy the product for the company to hit a certain revenue threshold. And so what'due south the optimum price that maximizes profit while minimizing toll? Microeconomics tin can respond that.
Macroeconomics, on the other hand, sees society at the macro scale. How technological innovation will change the economy? What is the bear upon of increased temperature on agricultural GDP? How global pandemic will hit unemployment? These are the kind of questions macroeconomists are typically interested in.
Domains of application
Y'all'll notice economists walking effectually in all domains — from unemployment and inequality to the economics of climate alter, to advertizing & revenue collection. Here is a small-scale sample of examples, much of which volition expect familiar to most data scientists.
Behavioral economists can quantify consumer response & mental attitude in response to marketing campaigns, cost reduction and improver of new features
Agricultural economists help decide how to ready the right insurance policy for farmers against natural disasters.
Industrial economists determine the level of article production and pricing machinery for a given market demand
Sports economists are already making a huge influence on " the way players are drafted or how much they are paid, through to individual coaching decisions, and fifty-fifty strategic shifts across entire leagues" [source].
Econometrics is the application of statistical modeling in understanding complex social and environmental issues. That's really a big area of applied economics, and I wrote a whole new article about it.
Methods & techniques
If non all, a large part of economics is data-driven and quantitative.
In quantitative economics, economists create a model of the real earth — a small or big part of information technology — in order to understand relationships between different components, their interactions and the impact of external influences (due east.g. policies, incentives, shocks).
And then, just similar the modeling of whatsoever other complex interconnected arrangement, economists rely on a range of mathematical and statistical techniques.
Mathematical models such every bit differential equations are used to quantify marginal changes in output for a small-scale change in input. For example, how a small decrease in price will drive the buy of a product and affect company bottom-line?
Statistical models, on the other hand, are used to develop and examination theories on the relationship betwixt economical agents (people, farms, businesses) and impacts of policies/incentives using statistical tools. In fact, the whole domain of "econometrics" is based upon the application of standard statistical models. Data scientists should be very familiar with well-nigh all of them:
- Descriptive statistics and hypothesis testing
- Correlation and regression analysis
- Time series assay and visualization
- Predictive analytics and forecasting
- Panel data modeling: OLS, fixed & random-effects models
As an bated, many economics questions are empirical, which tin can exist answered by post-obit the standard scientific processes— developing a hypothesis, collecting data, testing the hypothesis, making conclusions. All the same, economists can as well be too philosophical at times. They would often ask questions that are subjective in nature and the outcomes of which cannot be measured using objective metrics.
Analytics toolbox
If nosotros look dorsum 20 years from now, spreadsheet programs (primarily MS Excel) were economists' cardinal tools for building mathematical and statistical models. Excel was near exclusively used for visualization, and it is in use by a large pool of economists even today.
There too has been a modest but different breed of economists who liked programming languages for statistical modeling. STATA has been (and continues to be) the most popular programming platform for economists. A smaller portion of programmers would have other programs such as MATLAB and SPSS in their toolbox.
Things are changing speedily with a new generation of economists. They are embracing open-source software platforms more and more each mean solar day. Python and R are becoming the become-to platforms for many economists nowadays, although information technology is difficult to tell which one is more than pop, Python or R.
Many economists, especially macroeconomists like to play with time-series data for dynamic visualization and tell stories with data. For them, Tableau seems to be a groovy new toy to play with. I know a few of my economist friends, who nowadays utilize exclusively Tableau for interactive information visualization and creating web applications.
Parting words
With blurring disciplinary boundaries it is difficult to tell if a particular research trouble falls within a specific field. Numerous economists are now working as data scientists in different fields. Similarly, many information scientists are working with economic and behavioral problems which were exclusively the domain of economists in the by.
Every discipline is unlike only there are similarities as well in terms of bug solved, methods used, theoretical constructs and tools applied. So it'due south always skillful to pick upwards a few things here and at that place from other disciplines to strengthen inquiry & analytics chapters.
Statistical Tests You Need to Know in Economics
Source: https://towardsdatascience.com/economics-101-for-data-scientists-36cf6b3f0a15
0 Response to "Statistical Tests You Need to Know in Economics"
Post a Comment