Last Monday I introduced our most recent research project, involving Google search results and the impressions they give of Chicago's many neighborhoods. I outlined the goals and methods of the project there. I also revealed the 55 neighborhoods that are only mentioned by real estate companies on the first page of results, and explained why those results are unreliable when trying to choose a neighborhood to live in.
Today we're going to start getting into the heart of the matter. If you were using Google to research Chicago neighborhoods, would its results influence your initial impressions in a good or bad way?
Here at RentConfident we've got some experience in creating scoring systems. We designed our Confidence Factor as a means of weighing risks in Chicago apartments, and include it with every Signature Report. When it came time to analyze our search engine results I applied those same scoring design skills to create a system that tried to make something objective out of a very subjective thing. After all, can you really quantify the emotional impact of one adjective over another? It turns out that to some extent you definitely can.
Every neighborhood started with a score of 0. Here's the scoring that I came up with:
In the search results (titles and descriptions)
- Trendy/Hot/Hip: +2 points
- Other positive words: +1 point each (nice, pleasant, growing, renewal, etc)
- Shooting/Murder/Gang or other specific mention of criminal activity: -2 points each
- Housing Project: -1 point
- Other negative words: -1 point each (ugly, dirty, getting worse, fading, etc)
In the "related searches" area
- "Things to do": +2 points
- Hotel: +2 points
- Restaurants, Bars, Shopping: +1 point each
- Specific mention of criminal activity: -2 points each
- Crime or Crime statistics: -1 point each (yes, sometimes both appeared)
- Safety: -1 point each (as in, "Is [neighborhood name] safe")
The following words were noted, but not weighted:
- Mentions of race or economic class along with catchall words such as diverse, ethnic, immigrant
- Mentions of a specific lifestyle/culture (yuppie, hippie, hipster, gay, etc)
- Names of important people (Obama, Chief Keef, Walt Disney, Hugh Hefner, etc)
- Small/Large and similar
We'll discuss some of these in weeks to come.
Once the points were noted, each neighborhood's results were added up. Neighborhoods with a score of 2 or higher were marked as "positive". Scores from -1 to +1 were "neutral." Anything below -1 was "negative."
If you look at the point values we used above, each neighborhood by default had a chance to gain or lose 8 points if no duplicates were found. As it turned out several duplicates did appear, both in the search results and in the related searches suggested by Google. In the end we wound up with scores ranging from -17 on the low end to 10 on the top, with only 37 neighborhoods coming out indisputably positive. The pie chart below shows the breakdown of all the neighborhoods, excluding the "fake 55" listed last week.
The large number of negatives makes it seem like its tough for a neighborhood to catch a break from Google. Positive ratings were severely outnumbered. However, looking at the score breakdown it becomes clear that Google is actually quite neutral about the majority, with only some major outliers pushing into the far upper or lower reaches of the point spread. The bar chart below shows the actual spread, with the entire neighborhood list shown in blue and the "real neighborhoods" (fake 55 excluded) in red.
The neutral and low positive/low negative values were definitely favored. While it is tough for a neighborhood to get an extremely positive result, this is to be expected given that we're talking about Chicago, a city with a reputation for being a little rough on the whole. Additionally, once real estate sites are removed from the search results we're left with news articles and discussion forums. News articles are notorious for taking an "if it bleeds, it leads" mentality, and forum participants are also more inclined to gossip about bad things as a way of warning outsiders away from danger.
Given that tendency towards the negative, I was inspired to see what would happen if the neighborhoods scoring "+1" were seen as positive instead of neutral. There were 28 "real" neighborhoods that scored +1. Coincidentally, that's also the exact difference between the overall positive neighborhoods and the neutral ones.
In other words, if we take into consideration the slightly negative bias of our remaining sources, the pie chart included above flips completely to look like this:
The 10 best and 10 worst
Of course I'm sure you're all eager to know where your neighborhood stacks up in this whole affair. No cliffhangers this time. First off, the 10 best and 10 worst.
- Boystown (10)
- Chinatown (9)
- Andersonville (8)
- Lincoln Park (8)
- Wicker Park (8)
- Hyde Park (6)
- Lincoln Square (6)
- Magnificent Mile (6)
- Streeterville (6)
- The Loop (6)
There were no other neighborhoods tied at 6 points.
- Englewood (-17)
- Austin (-12)
- Riverdale (-10)
- Washington Park (-9)
- Fuller Park (-9)
- West Garfield Park (-7)
- Little Village (-7)
- Gresham (-7)
- Grand Crossing (-7)
- Golden Gate (-7)
There were no other neighborhoods tied at -7 points.
Of course, we did survey 234 neighborhoods, and that's just 20. Here's a map of all of them.
Is it Really that Drastic?
These results are, of course, unreliable. These results are a snapshot based on the day the searches were performed and the computer used to search. Current events, different search history, mobile vs desktop searching - all of these will change the overall results. Additionally, when viewing a search result page, you don't just see the individual words we used for scoring. You see a mishmash of sentence fragments and search terms, some underlined, some not. You also see all of the real estate results that we excluded as biased.
There's very few neighborhoods that scored unanimously positive or negative. There's even fewer where the real estate results were entirely forced off the first page. Most folks do trust real estate agencies, which put a positive spin on every neighborhood. Their presence in search results would in most cases buoy up the overall emotional impact of first page results to make a slightly negative neighborhood seem okay. It's only the ones at the bottom of the heap where the relentlessly chipper real estate results leap out as glaringly out of place.
I encourage you to search for your own neighborhood in a search engine of choice and see how the results (or any search engine's results) make you feel. Let me know if you uncover anything shocking.
Next week we'll be continuing this series looking at mentions of race, class and lifestyle choices in the results.
RentConfident is a new small business that protects Chicago renters from bad landlords. Do you want to see more posts like this? Do you want to see us succeed? Please share our posts with your friends using the social media buttons below! It helps more than you know!