Do you need data science training to build a linear regression model?Reading time: 3 minutes
I recently read an article in Medium that explained how to build a linear regression model using Python and Excel. After reading the article, I wondered if someone like me – someone without a background in data science – could replicate the outcome in Rapid Insight’s user-friendly predictive modeling application, Predict, and how much time it would require.
Our code-free (but code-friendly) tools are designed to make tasks like predictive modeling as straightforward as possible. I decided to put our tools (and myself) to the test and see how quickly I could match the article’s modeling outcomes.
Start the clock!
The author of the Medium article used a Boston Housing Dataset to create their model. I downloaded a copy to my local drive. Next, I opened Predict and loaded in the data.
30 Seconds in: A Bird’s Eye View
The first screen you’ll see after loading up a dataset in Predict is a high-level overview:
Our example dataset has 406 total records and 15 unique variables. From this view, you can assess the dataset’s readiness for modeling and quickly answer questions like:
- How many total records do I have?
- What variables are available?
- Are there any missing values?
This snapshot can help you identify any red flags before building your linear regression model. Thankfully, the Boston Housing dataset was clean and ready for modeling.
2 minutes in: Building a Model
After assessing the dataset’s readiness for modeling, you can automatically build a model using the ‘Model’ tab in Predict. In the Medium article, the author selected ‘MEDV’ (Median Home Value) as the outcome variable, so I did as well.
According to the article, the models created in Python and Excel used all of the variables in the dataset. I did the same, moving all variables into the “Included Variables” pane in Predict.
Next, I clicked the ‘Build’ to build the model. You can see that Predict built the model in a flash (0.03 seconds):
4 minutes in: Making a Prediction
Next, I created a .CSV file to store the values and used the model to predict the ‘MEDV’ price.
The model produced a new column (O) labeled ‘Predicted Y,’ representing ‘MEDV’ in this case. And just like that, I matched the predicted ‘MEDV’ from the Medium article: 19.41889. All in 5 minutes without the need to code anything!
Building Linear Regression Models without Data Science Training
Now that we’ve walked through just how quick and easy it is to build a model in Predict, let’s return to the question that inspired this post:
Do you need a data science background to build a linear regression model?
The answer: absolutely not, especially if you’re using Predict! The software makes it possible for users of all backgrounds to build predictive models.
Better yet, Rapid Insight users get free, unlimited access to our team of support analysts. If you ever run into roadblocks, or even if you just want a second opinion on your model, reach out and we’ll offer second-to-none support!
If you’re interested in learning more about bringing analytics in-house at your organization, click the link below to schedule a personalized demo!