Request
Hello,
I am student at*** and am on working on textual analysis using machine learning techniques under Natural Language Processing as a requirement for my master's degree.
I wanted to request for the text narrative data for crimes for the past 10 yrs. I have read the paper by Mr. Da Kuang on Crime Topic Modeling and want to do additional analysis on the efficiency of the methods that are being used for the crime topic modeling.
The data will only be used for Academic research purposes and here are some of the details that would be helpful in have in the dataset:
• Type of Crime as assigned by the LA PD
• Date of the reported crime
• Textual narrative of the crime event ( as reported by the police officer)
• Crime Code assigned
• Crime Category ( as assigned by the LAPD)
Additional Considerations:
1. NO PII requested.
2. ONLY CLOSED CASES ( I do not need data from any open cases/investigations)
3. Approx. 10 yrs of narrative text data.
Any other fields in addition to the above would definitely be useful but not necessary.
Formal crime types are not discrete in topic space. Rather, crime types are distributed
across a range of crime topics. Similarly, individual crime topics are distributed across a range of formal crime
types. Key ecological groups include identity theft, shoplifting, burglary and theft, car crimes and vandalism, criminal
threats and confidence crimes, and violent crimes. Though not a replacement for formal legal crime classifications,
crime topics provide a unique window into the heterogeneous causal processes underlying crime.
Keywords: Machine learning, Non-negative matrix factorization, Text mining, Crime
I do not need any identifying information and therefore all and any identifying information can be removed prior to sharing the data.
Kindly let me know if this is the right venue and method to request for such data and if anything else needs to be done in order to obtain this data.
With Regards,
Sincerely
I will be using the textual narratives taken down by the police officers to answer a couple of questions.
1. To understand the relationship between the behavioral and situational clues that lead to a crime event what is their relationship with the various crime types as classified by the formal crime classifications.
2. Formal crimes categories are distributed across multiple crime topics, and similarly individual crime events will relate to multiple formal crime categories when analyzed at the behavioral level. My analysis will help create new crime categories which would include the behavioral and situational aspect of the crime event as opposed to the current formal crime categories. This will not replace the current crime categories but help in providing a unique perspective towards understanding the reasons leading to a crime event.
3. I will also be using the metric of "perplexity" and "coherence" to evaluate the efficiency of multiple methods used for Crime topic Modeling.
Response
Thank you for contacting NYC Open Data Help Desk. Please be advised that narratives contain sensitive and private information including victims names, addresses and phone numbers. There are no plans to release this.
Thank you,
- NYPD
