Academic project analyzing a LinkNYC data using Python and Tableau. Applies Business analytics including Funnel, cohort, RFM segmentation to explore real-world business scenarios, uncover patterns, and generate actionable insights to support decisions for stakeholders
LinkNYC was created to offer free public Wi-Fi and digital access across New York City, aiming to reduce the digital divide. After several years of operation, the key question is:
Is LinkNYC a functional utility—part of the city’s essential digital infrastructure—or a fleeting convenience used only occasionally?
This distinction matters for future funding and policy decisions. To explore it, we analyzed weekly usage data (2020–2025) using activation, conversion, retention, cohort, and segmentation analyses to determine whether LinkNYC is used as a daily necessity or an occasional convenience.
- Objective: Assess whether LinkNYC serves as a long-term public utility (consistent, essential use) or a short-term convenience (occasional, opportunistic use).
- Key Insight (teaser): Our funnel analysis shows that 23.5% of total sessions reach the “High Value” stage, highlighting meaningful long-term engagement. Segmentation results reinforce this, with Engaged (64) and High Value (75) weeks representing the most consistent, high-impact usage, compared to Quality Sessions (122) and Convenient (39) weeks. This shows that only 13% of usage weeks align with convenient behaviour. Overall, these patterns suggest that LinkNYC functions more as a reliable public utility than a fleeting convenience.
- Methodology OverviewTo evaluate whether LinkNYC functions as a utility or a convenience, we applied several complementary analytical approaches:
- Session Analysis: Measured sessions per user and total activity to gauge overall engagement intensity.
- Activation & Conversion Metrics: Defined activation as weeks above the median data usage, and conversion as weeks in the top 25% of usage, to track progression through the user engagement funnel.
- Funnel Analysis: Built a behavioral funnel with four stages—Total Sessions → Quality Sessions → Engaged Sessions → High Value Sessions—to measure progression toward high-value, consistent use.
- Retention Tracking: Analyzed user sessions across seasons and quarters to assess persistence and habitual use.
- Quarterly Performance Analysis: Compared year-over-year trends (2020–2025) to measure growth, stability, and shifts in usage patterns.
- Cohort Analysis: Grouped users by seasonal cohorts to identify behavioral differences and long-term retention trends.
- RFM Segmentation: Classified usage weeks by Recency, Frequency, and Total Bandwidth (TB) to distinguish between convenience users and utility-level users. Together, these methods provide a multi-dimensional view of user behavior—linking engagement intensity, temporal consistency, and value generation to assess LinkNYC’s role in the city’s digital ecosystem.
-
Source: LinkNYC weekly usage dataset
-
Coverage Period: 2020–2025
-
Granularity: Weekly records of users, sessions, and TB downloaded
-
Key Metrics:
- Sessions per client
- Number of sessions
- TB downloaded
- Session length
- Data usage per client
- Derived metrics: activation, conversion, retention,
- Activation metric: A week is activated if TB_downloaded > median → measures engagement threshold.
- Conversion metric: A week is converted if TB_downloaded ≥ top 25% → identifies power users.
- Retention: Track sessions per client, Number of Sessions over seasons/quarters → assess persistence of usage.
- Why: These metrics collectively measure sustained vs transient interaction.
- Method: Aggregate usage(Number of sessions and sessions per client) quarterly from 2020–2025.
- Objective: Identify growth trends and volatility over time (YoY comparisons).
- Justification: Utility-like behavior should show consistent or growing engagement; convenience behavior would show decline or irregularity.
- Result Summary:
- Definition: Weekly usage grouped by seasonal cohorts (Winter, Spring, Summer, Fall).
- Goal: Identify seasonal retention patterns — do certain cohorts maintain activity longer?
- “Summer and Fall cohorts exhibit a more sustained session while we have high fluctuations accross years for Spring and Winter Cohorts
- Approach: Within the first quarter in 2025, track weekly session evolution.
- Purpose: Reveal short-term engagement or persistence pattern within a cohort.
- Interpretation: Gradual decline = convenience; steady use = utility.
Number of Sessions for each monthly cohort from week 1 to week 4
To assess how users interact with LinkNYC over time, we developed a four-stage engagement funnel reflecting the journey from casual to high-value usage. The stages were defined as: Step 1: Total Sessions – all recorded user sessions, representing initial access to the network. Step 2: Session Quality First, we defined quality sessions as those with an average session length of more than 5 minutes. We chose this threshold to filter out users who may have interacted with the kiosk out of curiosity rather than for meaningful use. Step 3: Engagement Next, we considered a session engaged if the user downloaded more than the median amount of terabytes (TB). This helped us identify users who were more actively using the kiosk’s capabilities. Step 4: High-Value Users Finally, we defined high-value users as those who met the engagement criteria and were in the top 25% (i.e., above the 75th percentile) of our data in terms of usage.
After completing the funnel analysis, we found that 23.5% of all sessions qualified as high-value. This indicates that LinkNYC kiosks serve more than just a momentary need, they provide real utility to a significant portion of users.
Conversion table
Overall Conversion Funnel
Criteria:
- Usage = TB Downloaded(TB)
- duration = Average session lenght (minutes)
-
Segmentation Logic:
Segment Criteria Interpretation High Value Week Top 25 data usage, Usage > 25% Core utility users Engaged Week Moderate–high use,up median data usage Regular users Quality sessions Light, intermittent, Duration > 5 Convenience users Convenient weeks Minimal use, all remaing weeks Opportunistic -
Results Summary:
- Quality sessions: 122 (dominant group)
- High Value: 75 ( 2nd dominant group, proving usage)
- Engaged: 64
- Convenients weeks: 39
- → Suggests only 13% of usage weeks align with convenient behaviour.
-
RFM Results Interpretation:
The analysis shows that most LinkNYC usage is driven by Engaged, High Value, and Quality sessions, with Convenient weeks accounting for only 13% of activity. Engaged users are the most recently active, while High Value and Quality sessions reflect consistent, high-bandwidth use. This pattern indicates that LinkNYC is primarily used as a reliable, habitual service, supporting its role as a functional public utility rather than a fleeting convenience.
Strong and consistent engagement among Engaged, High Value, and Quality sessions.
Low recency for Engaged weeks indicates ongoing, active usage.
High frequency and substantial TB usage for High Value and Quality sessions show meaningful, habitual interaction.
Seasonal patterns (Fall and Spring cohorts) demonstrate stable usage over time.
Convenient weeks account for only 13% of usage, highlighting limited opportunistic use. Lower session frequency and smaller data consumption in Convenience weeks indicate occasional, non-essential usage.
Key Insight: Overall, the predominance of high-frequency, high-bandwidth segments confirms that LinkNYC functions more as a reliable public utility than a fleeting convenience.
User identity: Sessions per client are treated as unique users, which may overcount usage from shared or public devices. The first three months of data recorded high session usage: Observed a spike in sessions in the early period of data, previous to COVID After COVID, sessions dropped and didn’t recover: Noted a permanent decline in usage post-pandemic, suggesting behavioral shifts or reduced reliance.
Temporal assumptions: User cohorts are defined by calendar seasons rather than event-driven. Conceptual Framing Assumptions: We made our own assumptions of what’s considered a utility: Defined "utility" based on usage patterns and criteria, not a standard created by LinkNYC.
- Leverage Utility - Level Engagement to Attract Advertisers
- Monetize Meaningful Usage Through Premium Ad Products
- Treat LinkNYC as an essential public service—invest in reliability, content quality, and local relevance to keep users returning.
- Use this analysis to show both commercial value (consistent usage = ad ROI) and civic value (public connectivity impact).
https://docs.google.com/presentation/d/16TzspxGnYVdPlgbSbmf9s45t-hRwqRBzpjMkg1IRt18/edit?usp=sharing