NYC 311 Complaints: Insights for a Better City

The city speaks, we listen.

Comment

Hackathon Project Story: NYC 311 Complaints Analytics

Inspiration

I recently visited New York City and, while exploring, I personally saw many problems in the city — from noise to traffic congestion and sanitation. It made me wonder: how is this considered the best city in the world if daily life issues remain unresolved?

Fortunately, I came across the NYC 311 dataset, which contains complaints filed from 2010 to the present day. This gave me the perfect opportunity to analyze patterns and validate what I experienced firsthand.

Due to a 100 MB limit in Salesforce Data Cloud, I was only able to use 15 days of September 2025 data. Otherwise, I could have covered many more years and explored long-term complaint trends.

How I Built the Project

1. Data Acquisition

  • Downloaded the NYC 311 dataset (NYC Open Data Portal) for Sept 1–15, 2025.
  • Reduced file size (<100 MB) to comply with Salesforce Data Cloud import limits.

2. Data Upload into Salesforce

  • Used the Salesforce Data Import Wizard to upload the dataset.
  • Selected a custom object called NYC_Complaints.
  • Mapped key fields during import:
    • Created Date → DateTime
    • Closed Date → DateTime
    • Complaint Type, Descriptor, Borough, Agency → Text
    • Incident Zip → Text (not Number, so leading 0’s aren’t lost)
    • Latitude, Longitude → Number
    • Resolution Description → Long Text
  • Dropped non-essential columns (e.g., Vehicle Type, Bridge Highway) to keep the dataset clean.

3. Building the Data Model in Salesforce

  • Defined a primary unique key (Unique ID) for each complaint to avoid duplicates.
  • Created relationships so that records could be queried across dimensions (Borough → Complaint Type → Status).
  • Set up metrics in Salesforce Data Cloud:
    • Complaint Count = COUNT(Unique ID)
    • Status-based metrics (e.g., Open vs Closed complaints).

This created a single-source data model that Tableau Next could easily consume.

4. Connecting to Tableau Next

  • Connected the NYC_Complaints object in Salesforce Data Cloud directly to Tableau Next.
  • Published the dataset as a data source.
  • Built visualizations:
    1. Top 10 Complaint Types – Horizontal bar chart, sorted descending.
    2. Complaints by Borough – Distribution snapshot.
    3. Daily Complaint Trend – Line chart (Sept 1–15).
    4. Borough × Complaint Type Heatmap – Highlights borough-specific issues.
    5. Agencies vs Complaint Types – Bar chart showing NYPD, HPD, and DOT workload.
    6. Donut Charts – For borough-level complaint proportions.

Business Value

  • Agencies can see which boroughs need urgent attention.
  • Policy makers can prioritize resources based on most common complaint categories.
  • Citizens gain transparency into what issues dominate their neighborhoods.

Challenges Faced

  1. Data Size Limitation

    • The 311 dataset is massive (millions of rows).
    • Had to filter down to 15 days to meet Salesforce’s 100 MB limit.

  2. Tableau Next vs Tableau Desktop

    • Tableau Next is quite restrictive compared to Tableau Desktop.
    • It took time to learn Salesforce Data Cloud and Tableau Next together, since I’m more used to Tableau Desktop’s flexibility.
    • If I had Desktop, I could have built richer visualizations, but this challenge taught me to adapt to constraints and still solve a real business problem.

  3. Field Mapping Restrictions

    • Couldn’t create custom fields during import.
    • Solved by pre-cleaning the CSV and aligning columns with Salesforce schema.

  4. Data Cloud Performance

    • Initial queries ran slow.
    • Improved speed by dropping unused columns and optimizing metrics.

Learning Outcomes

  • Learned how to connect open datasets into Salesforce Data Cloud.
  • Designed a data model from scratch with metrics and relationships.
  • Understood how Tableau Next + Salesforce Data Cloud integration creates a seamless pipeline.
  • Appreciated the importance of storytelling with analytics, not just raw dashboards.

Conclusion

This project shows that even a short timeframe dataset can uncover powerful insights about urban life. By combining Salesforce Data Cloud + Tableau Next, I built an analytics pipeline that can help city agencies prioritize resources, respond faster, and improve citizen satisfaction.

Built With

Share this project:

Updates