Real Estate Analytics

Powered by NoviLens

This case study explores how to use NoviLens to explore complex data, how to incorporate additional data to both solve a problem and to de-risk your decisions.

Start With a Question

Let’s assume that we want to invest in rental property in Los Angeles. A good place to start is to find a likely location for good deals. Sometimes, good deals can be found by purchasing foreclosed properties. The following question would lead us in the right direction: “Where are the most foreclosures of multi-family units in LA?”

To answer this question, we’ll need data. This foreclosure data is publically available from the LA open data project at This data includes location information and property type (single-family, multi-family, etc).

After the data is collected and uploaded into NoviLens, we create a dashboard with some widgets to see what the data looks like. We’ll add the following widgets:

  • Row count widget to see how our filters are working

  • Bar chart of # of foreclosures by zip code

  • A heat map of foreclosures

  • A pie chart of property types

  • A search bar to filter on particular zip codes

The dashboard looks like this:

Screenshot from 2019-12-10 11-08-15.png

No data is selected in the widgets, so we are viewing all the data in the dataset. Note that there are 3004 records in the dataset. Since we are interested in multi-family units, we’ll click on “Multi-Family” in the pie chart. The will change all the widgets to show only data related to Multi-Family units.

Screenshot from 2019-12-10 11-11-59.png

We can see from the row count widget that we’ve selected 405 records out of the 3004 record dataset. We can see in the map that the number of points are reduced and clustered in the center of LA. The bar chart shows which zip codes have the highest number of foreclosures.

Formulate Additional Questions and Add Data to Find Answers

It’s great that we were able to quickly filter the data down to what we are interested in, but this doesn’t really narrow down where to look for properties. The next question to consider is “Which zip codes have wage earners that can afford to rent from me?”. This question leads to a number of considerations:

  • What are the average wages by zip code?

  • What are the assessed values of properties by zip code?

To answer these questions, we’ll need wage data and assessed property value data:

  • Average wages by zip code is included in US Census data

  • Assessed property values are included in LA open data

Since the wages data comes from another dataset, we’ll need to do Data Fusion to relate this data with the LA open data. This is easy, just import the data table and relate it to the LA open data table by zip code.

Now that we have collected the data and fused it with existing data, we build a few more widgets:

  • A bar chart of average wages by zip code

  • A histogram of property values binned in 200,000 increments

  • A bar chart of property values by zip code

Now our dashboard looks like this:

Screenshot from 2019-12-10 11-35-57.png

Since we can’t afford more than $1,000,000, we can select all histogram bins less than $1,000,000. This will do a number of things. The assessed value per zip code widget will only show average values for all properties less than $1,000,000, and the wages widget will only show average wages in the zip codes represented by all properties that cost less than $1,000,000.

Since we want to eventually make money with this purchase, let’s make sure that the average wages can afford the rent. To filter the data, select all wages greater than $40,000 in the wages by zip code widget.

After filtering the data, the dashboard looks like this:

Screenshot from 2019-12-10 11-53-50.png

Notice how the points on the map are clustered in a fairly small area. This gives us what we need to call a commercial real estate agent to get started looking for properties.

De-Risk Your Decision

One final check in our analysis is to make sure that other property owners in the area aren’t making big investments in their properties so we don’t run into a large capital expense to keep up with the competition. This could be formulated into “What zip code has the fewest building permits?”.

To do this, we need building permit data which is available from the LA open data project. We’ll download this data and fuse it to our existing dataset by zip code. We’ll create a stacked bar chart widget that show the number of building permits by zip code that is stacked by the type of building permit. We’ll add another widget to show the total value of improvements by zip code as a check.

Now the dashboard looks like this (after selecting the zip code with the smallest number of building permits):

Given this assessment, we should feel comfortable with looking for properties within the map area shown above.

Answer your business questions in minutes, not months.

Learn how Fusion Analytics can help your organization.