Data Laundry

The Datalaundry website is software that automates the process of cleaning, searching and finding patterns in data. Registered users are allowed to upload their data set to datalundry.com, and save it in a library for subsequent access. Users may easily perform simple tasks such as searching for records with certain properties (specified by constraints on selected fields). Additionally, users may specify a dependent variable of interest, and possible independent variables, and DataLaundry will find patterns where the value of the dependent variable is predicted by the values of the independent variables. Example data sets are provided on the website for users to experiment with the available functions.

Example Data Sets

  1. Results of 6,602 NBA Games from 2021-2026, including scores, pointspreads, and over/under line. This dataset is useful for searaching and discovering situations when the pointspread has not been accurate (i.e. the home team or the visiting team beat the spread a disprortionate percentage of the games).
  2. Daily Stock Data for a selected group of popularly traded stocks, for 3 years, including date, ticker symbol, open and closing price, volume, and various computed variables, such as change in price from previous day, price change in last 2 days, price change in last 3 days, and in past weeks, and others. (Under construction.)

Scientific Method

It is generally agreed that rigorous science is needed to find truth in nature. Rigorous science is usually equated with implementing the Scientific Method. There is no official definition of the Scientific Method.

The scientific method has evolved and changed in many ways over the years. It was used more or less unofficially since ancient times.

In the 14th and 15th centuries, "natural philosophy" was born; by the start of the 17th century, it had become "natural sciences." It was during the 16th century that Francis Bacon popularized the inductive reasoning methods that would thereafter become known as the scientific method.

There are various definitions of the Scientific Method. According to Merriam Webster the scientific method consists of "the rules and procedures for the pursuit of knowledge involving the finding and stating of a problem, the collection of facts through observation and experiment, and the making and testing of ideas that need to be proven right or wrong." The typical steps in the scientific method are:

  1. Observation — Noticing and describing a phenomenon or a group of phenomena.
  2. Question — Asking a specific, testable question about the observation.
  3. Hypothesis — Proposing a tentative explanation or prediction that can be tested.
  4. Experimentation — Conducting controlled experiments to test the hypothesis.
  5. Data Analysis — Analyzing the results to determine whether they support or refute the hypothesis.
  6. Conclusion — Drawing conclusions based on the data and deciding whether the hypothesis is supported.
  7. Replication and Peer Review — Repeating the experiments and sharing results with the scientific community for verification and further testing.

This definition immediately raises the question about the third step: How do you find a hypothesis proposing a tentative explanation. Generally, this is understood to mean that scientists who study a particular field, have reviewed previous research, and observed various patterns in the data. A "pattern" means that the behavior of a particular variable (the dependent variable) can reasonably be explained by a set of conditions or restrictions on some other set of variables. In other words the behavior of the dependent variable can be explained by the values of a set of explanatory variables (usually called the independent variables").

The Datalaundry website is software that automates the process of finding patterns in a set of data. Users upload data to the website, specify a dependent variable, and possible independent variables, and DataLaundry will take over from there. Tools are provided on the website to perform various data processing functions, such as specifying restrictions on a variable, or defining a new variable based on ranges of a variable, and specifying labels for values of a variable. Moreover, elementary descriptive statistics can be easily obtained.