Exploring “Bullshit” Claims on Denver Crime
Dataset Sourced — https://www.kaggle.com/paultimothymooney/denver-crime-data#crime.csv
Denvers Crime Data |EDA|
(My EDA used to refute/amplify “Bullshit” claims about Denver Crime)
In our society communication is a key value that connects and informs one another. Unfortunately there is a plethora of misinformation, propaganda, and marketing tactics communicated to society but one critical pitfall is the spreading of bullshit. Philosopher Harry Frankfurt wrote an essay “On Bullshit” which exposes the prominent factors and concerns of spreading or conveying bullshit. According to Frankfurt, unlike misinformation or propaganda, bullshit “involves a kind of bluff a matter not of falsity but of fakery it is produced without concern with the truth, it need not be false”. From what I have learned throughout my Information Science studies I know exploratory data analysis (aka EDA) can be used to limit or further spread this concept of bullshit. Based on Frankfurt’s idea of “bullshit” I wanted to focus on a personal experience and apply EDA to further refute or amplify these ideals and claims I’ve been exposed to about Denver’s most criminally active areas since I have moved here two years ago.
To perform my own EDA on “bullshit” claims I obtained a data set documenting Denver Crimes. The data contained information on crime categories, dates, location of occurrence, etc. I planned on moving somewhere around Denver after graduation. People have told me that Colfax and other parts of Denver are dangerous crime riddled neighborhoods and I should avoid those areas. I used the Denver crime data to get a overview and visual outline of what areas crime’s were taking place and the types of crime there were. My main question was what neighborhoods around the Denver area have the most or least crime rates and what those crimes consist of. I used python to identify simple head, tail and call functions to ensure my data rows and columns were accurate with no NA values. Thankfully my data was clean and did not contain any NA values to troubleshoot and clean. Once I knew my data was clean and reliable I started exploring the overall offense accounts per crime and then per neighborhood. My raw count results indicated that Denver’s highest crime offense was traffic accidents, and the lowest was theft from yards. The top neighborhoods with the most criminal offenses recorded were Five-Points, Stapleton, and East/West Colfax. The least criminally active areas around Denver were Indian-Creek with only 502 reports, compared to Five-Point’s 23,571 accounts. After gaining insight from the raw number counts I wanted to produce a visual representation of the distribution of offense types given their main category. I rendered a pie-plot that exposed the distribution of traffic reports compared to more serious crimes like theft, drug use and aggressive offenses.
After getting a supporting visualization rendered form the raw offence counts I saw some high rates of criminal activity I wanted to further explore to uncover where these hotbeds of specific activity were taking place. The first criminal offense I analyzed further was drug use, specifically methamphetamines so I could pinpoint those areas to avoid. To do so I made a new data frame that contained only methamphetamine crime occurrences and plotted a new visual to analyze with meth incident counts across different Denver areas. From my analysis the highest meth activity is in Denver’s Central Business District, Civic-Center and West-Colfax. The second criminal offense I looked into was murders which had very low occurrence rates compared to other crimes but it’s serious nature was of interest. The most murder reports stemmed from Montebello with 17 accounts, Five-Points with 16 and East-Colfax with 12. After analyzing some of the more serious crimes and personal dangers I wanted to explore traffic accident prone areas of Denver. I looked into high accident areas beyond other criminal activity because I am not only worried about my physical safety but the safety of my property like my car. After plotting the traffic occurrence only data it was clear the highest occurrence of traffic accidents occurred in Denver’s Stapleton, Barker and Five-Points areas.
Based on my process of exploratory data analysis I was able to amplify peoples claims of dangerous areas and uncover “bullshit” claims of what makes that area dangerous or unattractive to move into. From my data exploration and analysis I continued to come across disruptive criminal activity around Five-Points and the Colfax areas. Due to these outcomes I could effectively amplify the claims I heard about not moving to Colfax and other specific inner Denver areas. It was clear that most of the criminal activity was taking place in the heart of Denver. Due to the outcome of my data exploration and analysis the most practical take away for anyone searching where to move if you are concerned with vandalism of property and fear of personal harm would be to live on the outskirts of Denver’s city limits.
After presenting my current EDA process and findings I gained essential peer feedback on how to make my results more accurate using normalization. For example I could have normalized all the areas/neighborhoods categories using secondary population datasets. Using my current crime data with Denver area population statistics could give my analysis a more precise outcome since it would take into account high volume human activity versus private neighborhoods with low population. Taking population into consideration would be essential since my current results may be a bit skewed due to the higher probability for populated areas to gain more criminal reports. While one area may have larger number of criminal offenses it may be less dangerous than a area with low population with a high percentage those few people being criminally active. If I get more time in the future I can find and take into account the population counts for my data analysis results.
ADDITIONAL READ IF INTERESTED
“A penny tossed off the empire state building will kill someone it hits below.” Truth or Bullshit?
(Past EDA used to refute “Bullshit” claims in the wild)
Further examples of this idea of “bullshit” in the wild results in things like people having an unconditional fear of being eaten by sharks. This common fear has been manipulated by media and Hollywood perceptions. While many people are afraid of sharks and perceive the animals as aggressive there have been data collection and EDA that show shark attack rates are low to refute the impractical fear of that animal. I recall an article discussing this fear of sharks being ridiculous because statistics show more people are killed by vending machines falling on them than shark attacks each year. The author concluded that based on data and analysis people should be more afraid of vending machines than sharks. Unfortunately what society sees and are told about sharks being aggressive produces the imperative fear of being attacked by one. While I thought this example would have been a good one to do my own EDA on I don’t think I would source the required fatality data for vending machines in my assignment time constraint.
An additional example of “bullshit” revolves around the common tale told to generations that if a penny was dropped from the empire state building it could gain enough velocity to kill a pedestrian standing below. I always thought this claim was practical based on the solid composition of a coin and the distance it would fall from allowing it to gain enough speed to seriously injure and kill a human who was hit. While the idea seems practical the earth’s characteristics like air density, gravity and other physics play a large role in refuting this bullshit claim. To prove the bullshit surrounding this claim I thought about how an object would react differently given a real world experiment versus a vacuum chamber where earth’s features like air density, and wind resistance don’t exist. Beginning EDA to refute this understanding I questioned what the severity of harm would be produced by different falling objects like a flat coin compared to an object with more aerodynamic qualities like a dart. While I found it hard to source a dataset around velocity of objects based on size and weight I did source statistics, facts, and velocity experiments that I was able to reference that proves the bullshit around this fatal coin drop tall tail. If a penny was to be dropped from the top of the empire state building it would gain accelerated speed due to gravity pulling the object downward. On earth the gravitational force can affect objects velocity based on shape, weight, and distance fallen.
Physicist Lou Bloomfield from the University of Virginia performed his own experiment on this tale of deadly pennies. Bloomfield conducted a mainly physical observation experiment without much reporting of raw data. He used wind tunnels and helium balloons to produce the appropriate velocity a penny would travel while using his own body as a data collecting sensor. Bloomfield tossed pennies into a high velocity wind tunnel and observed how they travel through air and what they felt like during human contact. The balloons he sent into the air full of pennies and had a drone pop the balloon when it reached required altitude. Bloomfield’s findings were both comical and reassuring. Bloomfield lived to tell his results around air molecules slowing down the penny as it falls which heavily restricts the penny gaining rapid velocity during free fall. He also noted pennies in flight tend to lay flat causing great air resistance and they would flutter back and forth. Bloomfield also results that pennies reach their terminal velocity only after 50 feet of free fall, then the two forces mass and air work against each other retaining a semi constant velocity all the way down. Meaning if you toss a penny off the empire state building it would drop 50 feet gaining around 25 miles per hour then begin to flatten out and flutter to the ground. Funny enough Bloomfield claims the pennies hitting him and falling on his head only resulted in a feeling much like being flicked hard in the forehead, nothing intense.
