Craigslist is one of the most visited websites in the Rio Grande Valley. Alexa and Quantcast both place Craigslist as a very popular site for people with an IP address from the counties of Hidalgo, Starr, and Cameron. Thousands of people in South Texas post classified ads in this free service and this provides a lot of open data for analysis. We decided to explore this dataset to find online behavioral patterns and help our clients better understand the Rio Grande Valley’s free real estate market.

Relevancy of Craigslist Real Estate Listings:

Craigslist disrupted the traditional advertising model of most print publications and even online magazines. According to a New York University study, Craigslist significantly affected the market for many local newspapers (Seamans, 2010). In the field of real estate, Craigslist has been a tough contender for most real estate brokers across the United States. The classifieds site allows home owners to post their offering on the platform and interact with prospects directly, this model significantly affects the traditional MLS distribution and attribution economy (Hahn, 2006). There is plenty of data in Craigslist that could benefit both real estate brokers and the general public. The language structure, posting time, and general geographical keyword orientation can provide valuable insight in very competitive market (Teece, 2010) .

Methodology:

Mining data directly from Craigslist using a crawler is against their terms of service. Instead, we used a public Google API that extracted data from Google’s cached (local copies of web pages) resources. The dataset included 2,300 posts in the Real Estate section. There were about 400 duplicate posts. Analysis takes into account classified posts from Feb 15, 2015 to March 1, 2015.

Results:

In general the most active days of posting classified ads for real estate in the Rio Grande Valley are Monday, Wednesday, and Tuesday. Most people posted between 10AM-12PM, there is a small increase in posting activity at around 3PM. The two most popular categories for people to post are “Apartments/housing for rent” and “Real estate by owner”. Together they make up 66.9% of all postings for the given time. People posting about “seeking an apartment for rent” expect to pay, on average, $400 per month. However, the average rent for most apartments posted on Craiglist is $636. To assess the general sentiment of the posts, we used an English lemmatization dictionary and the Roget Categorization Dictionary. We also manually translated posts written in Spanish and substituted words that included decorative/special characters. In general, most posts have a positive title. 38.7% of the posts can be classified as having a positive leading tone, 10.4% as having a negative one, and 51% as having a descriptive or neutral tone. The most common phrases in post titles are: For sale, For rent, for more info, house for, real estate, your credit, by owner.

Further analysis

It would be interesting to explore if the posting language or the usage of special characters in post titles has an effect on people clicking on the headline. In the future, we will be exploring the average life-span of a post and we will attempt to measure their success in actually helping homeowners in the RGV sell or rent their properties.

References:

Seamans, R. N. (2010). Technology Shocks in Multi-Sided Markets: The Impact of Craigslist on Local Newspapers. Hahn, R. W., Litan, R. E., & Gurman, J. (2006). Bringing more competition to real estate brokerage. Real Estate Law Journal, 35(1), 86. Teece, D. J. (2010). Business models, business strategy and innovation. Long range planning, 43(2), 172-194.