Introduction
In this document, I perform the real estate area assessment with a forward-looking lens. There are three reasons:
Mission Statement — The mission statement of the college is to Produce career-ready, global business leaders through hands on discovery and application. I believe that since all areas assess with exams; the assessment is not congruent with the mission statement. For example, a natural attribute of a career-ready graduate is to work well in teams. Exams do not measure team work.
Forward Looking — Upon a cursory review, prior assessments are reactive — they are not proactive. For example, the assessment says things like: this is what we did, and we were great at it. But most assessments do not say that this is what we will do.
My, Myself and Pratish — Unlike other concentrations, I am the only professor (Hamed is not fully involved yet) who handles everything real estate. Such power comes with both positives and negatives. An example of a negative is that I cannot take a sabbatical. An example of a positive is that I do not have to worry about coordination or hurting anyone’s feelings. As a result, this sort of experiment is easiest for me to pull off.
Considering the three reasons, I assess by analyzing real estate jobs on indeed.com. My analysis comprises two parts. In the first part, I scrape all full-time jobs in California with the keyword real estate. (https://www.indeed.com/jobs?q=real+estate&l=California&jt=fulltime&start=10). It turned out real estate agent jobs dominated this search. While useful, it does not require a Cal Poly degree to be an agent. In the second part, I consider refinements. I refined the search by scraping jobs with keywords: real estate acquisition and real estate development. Indeed.com also seems to have banned me from scraping — the cat and mouse games have begun. I chose the refined keywords after consulting with a couple of students who are in the job market.
After scraping, I using Natural Language Processing (NLP); I answer the following questions:
- Using the job position, what positions are the most popular for real estate students?
- Using the job summary, what are the most frequent words?
- Using the job description, what words are most correlated with skills mentioned above?
Note some caveats. The search for keywords is more of an art than an exact science. Also, NLP requires implicit assumptions, and hence the results may depend on “data cleaning”.
To be clear, assessment is difficult. On one hand, it is incumbent to teach analytical skills that are independent of jobs. For example, I do not think that program reviews should depend on whether we are in a recession. On the other hand, a program also needs to be relevant. So, if jobs require knowledge of Python, then a program should phase out Matlab. My primary goal is to do something different, something that is forward looking and something that is objective.
Main Findings
| Real estate license |
May be |
I accept real estate license as a senior project. I make it clear that I do not like it I will de-emphasize my displeasure |
| Communication Skills |
No |
Add a report section to the final project, Make students present, teach them how to make a report and presentation that looks good |
| Exams |
May be |
Exams are useless. Go toward a project based approach that emphasizes working in teams, presentation |
Note that ESL students are screwed. Most jobs require good presentation and writing skills; they will be left behind. Also, making students take a writing course is punting. Anecdotally, it seemed that most of students thought GWR was a waste of time. Some of my more technical students, actually, failed the exam once which again shows that they did not care for the exam. Writing needs to be incorporated in a more natural way rather than a course
Last, I always hated exams and it could be that I am finding things corroborating my hate. I suffer from confirmation bias. I will keep thinking about this issue.
First Part — Analysis of the keyword: Real Estate
I performed the scraping using the rvest package in R. First, using the keyword “real estate”, I scraped 99 pages of full time jobs in California. An example of the scraped webpage is https://www.indeed.com/jobs?q=real+estate&l=California&jt=fulltime&start=10. In total, I scraped 1533 listings. For each listing, I obtained the title, location, company, summary and description of the job.
Table 1: This table shows the first three listings with real estate as the keyword. I do not show the description for expositional clarity
| Real Estate Listing/Buyer’s Agent – New or Experienced |
Keller Williams/CA Realty Training |
Pismo Beach, CA |
Represent buyers and sellers to start and close real estate transactions. |
| Real Estate Agent: Licensed / Unlicensed |
Keller Williams Realty |
Grass Valley, CA |
Dual Careers - Get your Real Estate License while working your current job!! |
| Real Estate Sales Agent – New or Experienced |
Keller Williams/CA Realty Training |
San Luis Obispo, CA |
Represent buyers and sellers to start and close real estate transactions. |
Consider the first listing. A company named Keller Williams is looking for a real estate agent in Pismo Beach. The third listing is identical to the first except for the fact that the location is in San Luis Obispo. The summary of the second is listing is noteworthy as well. The summary mentions that potential employees can get their license and the company may pay for it.
To work with the job titles, I do some data cleaning. For example, I first remove the digits, spaces and punctuation. I also remove the words “real” and “estate”. Last, I transform all the words to be lower case. These cleaning steps are normal although the sequence matters. I performed different permutations of the steps to get a visual feel for the results.
After cleaning, I create a Document Term Matrix (DTM). Table 2 is an example where 12 cleaned words form the columns of the DTM and the listings form the rows. For example, the first listing has the word “agent” while the second listing has words “agent” and “licensed” in them.
Table 2: This table shows the words associated with the first three listings
|
agent
|
analyst
|
assistant
|
associate
|
commercial
|
consultant
|
coordinator
|
leasing
|
licensed
|
loan
|
manager
|
mortgage
|
|
1
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
|
1
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
1
|
0
|
0
|
0
|
|
1
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
0
|
The DTM forms the core of the NLP. Figure 1 shows the most frequent words from the listings. Two observations are in order. First, the word “agent” occurs 281 times which is 18.33% of all the listings. This is extraordinary. Said differently, 18.33% of the listings are for real estate agents which does not even require a college degree. The presence of “licensed” corroborates the finding. Second, the presence of keywords “mortage”, “loan”, “underwriter” implies that there may be a demand for loan officers as well. Again, a college degree may not be required to perform such a function.
To summarize, while it is instructive, I do not feel that the key word
real estate does a good job of figuring out potential job oppurtunities for Cal Poly students. Certainly, I do not want to assess my students based on Figure 1 implications.
Analysis with refined keywords
Analysis of job titles
In this section, I analyze the job description from the two sets of keywords: real estate acquisition and real estate development. In total, there were 3011 listings albeit some of them were duplicates. After removing the duplicates, I work with 2051 listings.
According to Figure 2, the story from the refined words is different than before. Real Estate agent is not that common anymore. This makes me feel good about the use of the keywords.
Analysis of Job Summary
Figure 3 shows the most correlated words with the word development. The x-axis shows the correlation. For example, the correlation between development and entitlement is more than 50%. This means that half of listings contain the word development and entitlement together. Other noteworthy correlated words are infrastructure, planning and construction.
Two observations are in order. First, it may be that development jobs are not meant for OCOB students. Some of the jobs may be for CM or CRP majors. The second point is related to the first. A real estate education should encompass knowledge related to entitlement, infrastructure, planning and construction — real estate should be more cross collaborative
Next, I focus on jobs that are explicitly finance focused. To do so, I filter for job titles that containing the terms: “accounting”, “analyst”, “asset”, “business”, “commercial”, “financial”, “investment”, “leasing”. All the terms are from Figure 2.
Figure 4 shows a word cloud that shows the most correlated words with “skills”. Loosely speaking, I interpret this word cloud as the most sought after skills from undergraduates looking for real estate finance jobs. The font size is related to the correlation. Two observations are in order.
Communication related skills are most important. This is corroborated by the presence of words like communication, written, and teams.
Some technical skills are related to analytical ability presented by either word or excel.
Figure 5 goes toward what is traditionally considered real estate finance. Some of the financial jobs that a student may do includes work supporting budgeting, forecasting (pro-forma) issues. The student may also assist in issues related to performance measurement, dealing with lenders, making financial statements and cash flow related issues. Knowledge of excel and word also shows up here.
Conclusion
In this Memo, I attempt to come up with an assessment of “real estate” using jobs. The idea is that I want the students to be career ready and I want to emphasize (deemphasize) certain skills. I find that an exam based assessment misses the point. I also find that I am not emphasizing communication skills as much as I should.
