The Kwelia Blog Has Moved!
We’ve migrated our blog to https://blog.kwelia.com/.
We will leave all the current posts here for now so that old links still work, but for new content please visit the new blog.
We’ve migrated our blog to https://blog.kwelia.com/.
We will leave all the current posts here for now so that old links still work, but for new content please visit the new blog.
Among apartment industry insiders, it is hardly a secret that Washington DC has been among the most impermeable markets since the housing bubble burst. Considering that it’s home to the most recession-proof business on the planet (the Federal Government) and steadily teeming with young, diverse talent, this shouldn’t come as much of a surprise. Beyond heavy demand for existing apartment stock, this market has seen a crazy new construction boom. While the City Center project has received much of the national acclaim, there are several, several other new construction projects under development. (see here).
Basic economics should help us conclude that these factors (resilient local economy, new construction, etc.) would lead to steadily increasing rents. Put differently, stable demand and limited supply (hence the new construction) begets increased pricing pressure. Despite this dynamic, however, recent reports have indicated the opposite effect. As a matter of fact, one major report indicated that rents on class A (nicest with amenities) and B properties have actually decreased by a whopping 8% over the past year. To dig deeper into these theories and substantiate some of these reports, we looked to our platform for some answers.
Increased Median Price Per Feet
This is a screenshot from our Competitive Intelligence product, which shows an exponentially smoothed time series of the median price per square foot of all DC apartment listings we could get our hands on in this timeframe (extensively filtered, of course).
According to this trend graph, rents have increased by 9% on a square foot basis from July of last year until now, moving from $2.43/ft to $2.65/ft. So, it is safe to conclude that rents definitely have not gone down in DC, despite what some reports have suggested.
A Deeper Dive
But while the city-wide data is helpful to see some macro trends, anyone that tracks real estate knows that metro-wide reports can be misleading. This is because such “broad-stroke” analysis can jumble together apartments of different types, classes, and locations, thus leading to bias. After all, like politics, real estate is inherently local, isn’t it?
To try to eliminate some of this potential bias, let’s see these rent price trends on a more local level. To help visualize this information on a more granular level, our Chief Data Scientist put together a heatmap showing rent price movements on a census tract level. For how he was able to put this together, see here.
Non-Uniform Rent Growth
Well these movements are rather interesting, aren’t they? According to the legend on the right, blue shows positive change, while red shows negative change, with the darker colors indicating the gravest changes. From this map, we can start to see that the growth in rents has not been uniform and has in fact been fairly disparate depending on the neighborhood (census tract).
Top Neighborhoods for Year on Year Growth
So where has the growth been the strongest? From the map, certain neighborhoods like Woodley Park, the Columbia Heights/U Street area, Capitol Hill, and the Brentwood/Catholic University area jump right out. Each of these neighborhoods had rent per square foot growth of somewhere between 27% and 30% in the past year.
So Why the Growth Disparity among Neighborhoods?
So what gives? Why have certain neighborhoods seen such tremendous rent growth, while others have not? Well, to answer this question we only need to revisit the first paragraph of this post – new construction. “Cranes are the most obvious signs of economic activity in the District today,” according to a recent WaPo article. In 2011, developers broke ground on nearly 15,000 new residential units. Late 2012 saw another 6000 units come live, which dwarfs the 2500 that New York City saw.
Increased Listing Volume
We have noticed this trend internally as well. From last year to this year, we have seen 25% more listings come through our platform, which has to be a direct result of the new construction boom.
What do these four neighborhoods (and others as well) have in common? They’re all hotbeds for new construction activity. While we’re not DC locals, it seems that some of these projects have gone live recently, thereby positively affecting rent prices.
More to Come in Future Posts
There were so many interesting insights from all of this data that we couldn’t fit them all into one post. Stay tuned for more DC market analysis.
Our last post which cataloged Philly’s most expensive places to rent in May got a lot of attention around the local real estate blogosphere (see here and here for examples). Our approach was to use rent price per square foot to rank neighborhoods’ relative rental costs.
While this approach is great for comparing neighborhoods to one another (look out for another post for June’s top 10 coming soon), it’s not as useful to individuals who are interested in understanding how much they can expect to pay for a particular unit in a particular neighborhood.
With that problem in mind, we created a new heatmap (shown above) where you can visualize the median rent price for a certain type of unit across neighborhoods (using bedroom and bathroom sliders to adjust the map values.) This map is updated daily to reflect the newest listings coming on to the market. Blank areas on the map reflect areas where we don’t have a enough data in the last 90 days to give a value.
Lately, it seems as if neighborhoods have been all the rage here in Philly. Recent blog posts have been indicators of this trend. There was this one in Philly Mag that identified the hottest neighborhoods in Philadelphia based on home sale velocity. Then there was this one in Curbed Philly that highlighted up-and-coming neighborhoods according to the sentiments of RedFin agents. What was most interesting about these posts, was that the top neighborhoods were ones that we would least expect. So, although some of these lists included the “usual suspect” neighborhoods, there were several sleepers. For example, Phoenixville, Brewerytown, and West Germantown were the top three up-and-comers according to agent sentiment at RedFin. In the Philly Mag piece, Graduate Hospital (of all places) was dubbed the hottest.
As data geeks, these articles got us thinking. If Phoenixville is hot according to agent sentiment and Graduate Hospital is hot according to home sales, where is hot according to apartment rental pricing? Considering that we track this kind of stuff, we figured we would find out by ranking Philly neighborhoods according to apartment rental pricing. Some past experiences have indicated that heat maps are a great way to visualize pricing trends of different geographic areas. With that in mind, here are the results for the month of May (click image to get to interactive version of it):
Who’s Number One?
At a whopping $2.41 per ft2, Fitler Square was the priciest rental neighborhood in Philly for the month of May. While the neighborhoods of the top 10 are mostly the “usual suspects” as referred to above (Rittenhouse Square, Old City, etc.), this was a mild surprise. As for what could be behind this, we have a couple of theories. The first is what we’ll call the “Naval Square effect”. For those not aware, Naval Square is an upscale megaproject by Toll Brothers. Despite being a condo project, owners there list their units for rent from time to time. Now although Naval Square is technically a part of the Graduate Hospital neighborhood (according to our neighborhood shapes), it does border Fitler. This could mean that some Naval Square price observations are getting included in Fitler, which would give it a nice price boost. The other theory is that since Fitler has a high ownership rate, the quality of units there may be higher than other parts of town. We’ll continue to monitor Fitler and tweak things like the minimum number of units per neighborhood. Stay tuned for updates in a future post.
Our Up & Coming Neighborhoods
Although outside of the top 10, we thought it was worth noting that a couple of the Germantown neighborhoods are fairly highly ranked on the list. For example, Germantown – Westside came in at number 14 and West Central Germantown came in at number 18 for the month of May.
Though surprising to some non-Philly/Germantown folks, this may not even be a fresh trend based on other charts we have. Both of these neighborhoods have more or less held steady near $1.50 per ft2 for the past several months. One does wonder whether these neighborhoods are uniformly hot or whether certain complexes are single-handedly propping up rates? In otherwords, are a few upscale complexes responsible for a higher-than-normal median rent? A closer look at complexes like Rittenhouse Hill, or Delmar Morris, or Cloverly Park shows that median rents there are all well north of $1.50 per ft2. We’ll take a deeper dive into these neighborhoods and report back accordingly.
Leasing Season Is upon Us
As most real estate pros will acknowledge, May is actually the beginning of “leasing season”. From now until September, listing volumes and rental prices will be at their highest points for the calendar year. When considering academic calendars and weather patterns, this rhymes with common logic.
There are a couple of neighborhoods affected by leasing season that we thought it noteworthy to mention.
The first is University City, which went from number 5 in the (unpublished) April rankings ($1.86 per ft2) to number 12 in May’s rankings ($1.62 per ft2). Students typically sign and make new leases around this time of year, so it follows logically that apartments marketed towards students might be higher in April than May. We will keep you posted in how this neighborhood continues to trend.
The second one is Society Hill, which went from number 19 in April to number 9 in May. Like University City, seasonality may be a culprit here. In addition to a strong rise in median rental price, we saw more than double the number of listings in Society Hill from April to May. This is perhaps an indicator that a glut of leases in this neighborhood reset during this leasing season. We shall see.
Neighborhood Shapes - We were able to cluster our data in neighborhood shapes thanks to our friends at Azavea. We were certain to only include neighborhoods that had a minimum number of price observations, which might explain some holes in the map.
Unit of Measure - The unit of measure here is the median rental price per square foot per neighborhood. Although imperfect, we felt that the price per square foot was the best way to capture price levels across all unit mixes. In other words, it was the best way to do an apples-to-apples comparison across all neighborhoods because one may have a lot of one bedrooms or studios, etc.
Bias - Lastly, note that we were careful to minimize any bias that outliers would cause by using medians and not averages of pricing data.
In true startup fashion, we will continue to iterate our methodology to produce the best results for this. Stay tuned for future posts, as we will be reporting on neighborhood pricing movements on a monthly basis. This should get interesting!
In our last post, our Chief Data Scientist plotted average rental prices per square foot in each ZIP Code in the San Francisco Bay Area through a quick hack using new mapping packages for R-Studio. In true startup style, we have iterated on that and now have something better based on the feedback that we received.
How have we iterated you ask? Well, we have iterated in three major ways:
First, we have made the map web-based and interactive. Now you can “mouse” over different areas to get information instead of tediously matching areas with our raw data. Also, in response to some of the comments, it is now easier to see different cities and towns underneath our pricing information.
Second, in tune with our quest to provide the most granular information possible, we have mapped the data by census tract. According to the Census Bureau, Census tracts are designed to be “relatively homogeneous units with respect to population characteristics, economic status, and living conditions… census tracts average about 4,000 inhabitants." We look forward to leveraging the public data available from the US Census to analyze things like income and demographics compared to rent in the near future.
Finally, whereas we presented the average price per square foot the last time, we presented the median price per square foot here. We have discussed this in a previous post, but medians tend to be a more telling depiction of pricing than averages because there is a lower likelihood of an outlier skewing the sample.
As we noted in our last post, it is always somewhat astounding to see the disparities of rent pricing in seemingly close geographic locales. But even more basically, it is incredibly eye-popping how high rents are generally. Then again, this should come as no surprise, as according to David Crowe of the National Homebuilders’ Association, “[a]ll of the net addition to households since 2004 has been in rentals.”
To play with the web-based map, please go here. Let us know what you think.
We created Kwelia because residential real estate information is unstructured, messy, informal, and overall - not helpful. A glaring example of this problem is the lack of granularity of real estate data on rentals in urban markets. While it is not difficult to find indexes and other metrics on pricing movements in a city level, it is difficult (if not impossible) to find such information on pricing movements within a city. Good luck finding pricing information per ZIP code, or better yet, by neighborhood.
Well, consider this problem solved. After tinkering around with a mountain of fresh rental data from the SF Bay Area, Kwelia’s Chief Data Scientist Chris Connell was able to plot the average rental price per square foot in each ZIP code on a map. He describes the technical details for how he put it together in R on his blog. Check out the map below or an enlarged version here:
It’s always interesting to view things geospatially. One key takeaway (besides the obvious one that the SF Bay is PRICEY!) is that there is incredible price disparity among different parts of this massive MSA. Although it is obvious that the East Bay is cheaper than the West, it is insightful to see that is it nearly $2/sqft cheaper. For further granularity, check out below for the actual raw data that was used for the map (enlarged link and raw data).
Can you find your ZIP?
While we here at Kwelia spend our days (and several nights) working to bring cutting-edge techniques in data science to residential real estate, every now and again, other interesting applications of our techniques arise. It’s fall of a presidential election year, so much of America is preoccupied with the impending presidential election. Few would disagree that the most entertaining components of the candidates’ campaigns are the debates. It’s always a good time to watch them verbally joust against each other to solidify positions on issues and manifest their campaign rhetoric.
Who is the Winner?
But although the debates can be fun to watch generally, whether to poke fun at a candidate’s hair or to yell and call another a liar, they tend to get frustrating because there is often so much dissonance as to who addressed a topic better or even who won overall. While the networks determine debate victory by polling citizens, there is typically crazy variance among the networks. This variance phenomenon was amplified during last Thursday’s Vice Presidential Debate. While no one disagrees that it was a close battle, who is to say (objectively) that one candidate completely pummeled the other candidate?
Well, this was what different networks told us according to their polling. According to the MediaMatter.org blog, “Snap polls released after the debate last night were mixed; a CBS poll of undecided voters found Biden winning 50-31, while CNN declared watchers “split” after their snap poll reported Ryan narrowly winning 48-44.” If this wasn’t biased (or utterly confusing) enough, the different sides are pointing to different (unscientific) polls as indicators of their side’s victory. For example, conservative media outlets are pointing to a CNBC.com poll that names Paul Ryan the winner – by a nose. But when you unsheathe the methodology behind the poll, it is nothing more than a popularity contest akin to that of a high school student government election. “Indeed, you can apparently vote multiple times across different browsers, and the results have fluctuated wildly over the past 15 hours. Last night, several conservative message boards and sites, including Free Republic and Tea Party Nation, posted links to the poll and encouraged their readers to vote in it. At various points, the poll has indicated that Ryan won the debate by twenty points and that Biden won the debate by 8. As of this writing, Ryan leads by 2 points with more than 190,000 votes cast.”
Let’s Find Another Way
So in order to decipher another way to objectively determine how the Vice Presidential fared through certain topics and even overall, our Chief Data Scientist decided to look beyond polling and analyze something more technologically forward…and even sexier. The answer was Twitter. His thought was that if you could measure the sentiment of all of the tweets transmitted during the debate, you could derive a fairly objective sense of what the sentiment is during certain topics. Further, perhaps it may be possible to aggregate positive sentiments and crown a victor as well.
Sentiment Analysis in Brief
For those that aren’t up to speed on sentiment analysis, Wikipedia describes it as analysis that “…aims to determine the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document. The attitude may be his or her judgment or evaluation (see appraisal theory), affective state (that is to say, the emotional state of the author when writing), or the intended emotional communication (that is to say, the emotional effect the author wishes to have on the reader).” Many of the techniques in sentiment analysis have been pioneered by renowned NLP professor Bing Liu. In fact, Professor Liu has authored together software packages that automatically parse words that determine positive or negative sentiment for tweets. These more or less set the standard for which words indicate sentiment. While analyzing tweets may come off as a simple exercise, it is instead rather cumbersome. Like our normal data routines, data must be collected, cleaned, and then ultimately presented in a format that facilitates further analysis. Please check out Chris’s blog for some insights into his process behind this.
The total data sample size for this experiment was 363,163 tweets, which was collected roughly every 60 seconds throughout the course of the debate. As we must do during out typical data collection work, we had to remove several tweets in order to clean the data. Duplicate tweets were removed, which left the final dataset tweetcount at 81,124 unique tweets whereby Biden had 52,303 tweets and Ryan got 28,821 tweets. Each point represents the series of tweets that were gathered each minute and intuitively, the farther above zero a point is, the higher the positive sentiment of the tweets (and vice versa).
Key Movements to Note
A quick analysis of the sentiment graph will demonstrate that there were some interesting peaks and troughs throughout the debate. We’ve gone through some of the most drastic ones to correlate it with what was going on in the debate when sentiments either rose or fell to such levels:
21:08 – This was during the foreign policy portion of the debate. You can note that Ryan’s sentiments were at lows during this early portion of the debate.
21:31 – This was during the piece when Biden accused Ryan of requesting stimulus funds. Ryan’s sentiments soared.
21:49 – This was during a Biden diatribe about what the Romney/Ryan camp may deem a small business (hedge funds perhaps?). Ryan’s positive sentiments soared again.
22:26 – This was during the closing statements for each candidate. Ryan’s negative sentiments reached lows.
Post-Debate - One interesting thing to make note of was the fact that although the debate only lasted an hour and a half, Chris was certain to continue the sentiment chart for an additional 30mins beyond the debate’s duration. As you can note above, there were some interesting movements for each candidate’s sentiment. It’s almost as if candidates’ respective sentiments were battling each other out for post-debate positioning.
While this exercise proved more tedious than expected, we were quite pleased with the outcome. If anything, it made watching the debates more entertaining. Stay tuned for more, however. Now that the code has been written, we will run the same analysis for the next two presidential debates – starting with tomorrow’s.
In advance of the impending Multifamily Revenue Management Conference, we were asked by UNITS Magazine to jot down some thoughts on some of the more salient trends in revenue management. For those not familiar, UNITS Magazine is the largest trade magazine for the multifamily industry. Please check out our section below. For the full article complete with the opinions of others industry professionals, click here.
Revenue Management: From Exclusivity to the Mainstream
For nearly a year, we have been hard at work on what we truly believe is the next generation of revenue management software for the multifamily industry. While developing our product, we have been able to witness widespread adoption seep through the industry as well as the emergence of some new trends. Most notably, however, we have observed that revenue management software (and the culture that comes with it) is no longer the exclusive domain of REITs and large corporates. As articles like “When Apartment Rents Climb, Landlords Can Say ‘The Computer Did It’” in the November 2011 New York Times will attest, Revenue management is slowly becoming the mainstream.
Enterprise Software Leads Revenue Management Software
Interestingly enough, the trends that have allowed for this movement to the mainstream have largely been a reflection of innovations in enterprise software as a whole. Over the past several years, enterprise software has undergone a disruptive revolution to allow it to adapt to rapidly shifting behaviors in the workplace.As such, we have seen things like cloud computing and customer relationship management systems become products businesses cannot live without.We are confident that revenue management software will approach ubiquity among multifamily professionals as it continues to take its cues from rapidly evolving enterprise software.Three trends remain prominent in this road to ubiquity: