Somantic - A visual search engine for real estate evaluation

I recently started working on a project called Somantic which will allow real estate investors to make visually intuitive rational decisions regarding property evaluation and the return on investment. The goal is to make Somantic easy to use and it will enable to filter the properties on the market which investors want to buy.

As the visualization I am using a self-organizing map which is currently the state-of-the-art to visualize and cluster multi-dimensional data that allows to examine non-linearity as well as providing a capability to preserve the topology and distribution of the data, which is very important in real estate valuation.

A self-organizing map (SOM) is a type of neural network that is used for unsupervised learning and data visualization. SOMs are often used for clustering and dimensionality reduction, and can help to reveal patterns and relationships in complex or high-dimensional data.

A simple example of a SOM is a map of a city, where each neighborhood is represented by a node on the map. The nodes are organized in a grid, and each node is connected to its neighboring nodes. The nodes are initially placed randomly on the map, but over time, they will move and adjust their positions based on the similarity of the data they represent.

For example, if we are using a SOM to visualize the demographics of a city, each node might represent a neighborhood, and the data associated with each node might include information about the population, median income, and education levels of the residents. As the SOM trains, the nodes will move and adjust their positions based on the similarity of their data. As a result, neighborhoods with similar demographics will tend to cluster together on the map, while neighborhoods with more diverse demographics will be more dispersed.

This simple example illustrates how a SOM can be used to visualize and analyze complex data in a way that is intuitive and easy to understand. By revealing patterns and relationships in the data, SOMs can help us to better understand and make sense of large and complex datasets.

(An example self-organizing map showing U.S. Congress voting patterns. Source: Wikipedia)

The idea for the project came to me last year when I started investing into real estate in Germany. I was searching (like everybody else) for best properties to buy which will rise in value and generate positive cash flow. Thus I needed a tool to do that.

I checked out websites like immometrica which I really liked to do some initial analysis but I was missing kind of the big picture of the real estate market in Germany.

The last push to actually build this service gave me Dr. Teuvo Kohonen himself since he unfortunately passed away on 13th December 2021 and I would like to demonstrate and make use of the underrated power of this visualization tool which he invented.

Teuvo Kohonen Source: Wikipedia)

You can find the latest version of the tool at somantic.net (!currently under construction! Alpha Release 0.0.2). If you want to receive updates regarding the status of the tool and be an alpha tester please send me an email and I will add you to the mailing list. For me it would be interesting to see how many people are interested in the tool, too. I am continuously improving the tool and I am glad to get feedback.

P.S.: Somantic is a mixed word between som + semantic.