Skip to content

speaker_notes.md

Latest commit

 

History

History
78 lines (64 loc) · 7.8 KB

speaker_notes.md

File metadata and controls

78 lines (64 loc) · 7.8 KB
Raw

🎙️ 2-3 Minute Hackathon Presentation Plan & Speaker Notes

This guide provides a high-impact, chronological script and interface action plan designed to nail your hackathon presentation in under 3 minutes.

⏱️ Timeline at a Glance

Segment Time Focus Interface Action Spoken Goal
0. The Hook 0:00 - 0:25 Welcome & Lihai's Editorial Start on the React landing page. Click "📰 Data Journalism Editorial" Ground the project in real-world civic impact.
1. The Platform 0:25 - 1:00 The 17 Cards & Filters Close tab, click "🖥️ Start Pitch Mode" Show how you consolidated messy drafts into a production platform.
2. Case Studies 1:00 - 1:50 Timetable, Heatmap, Poisson Click "Next ▶" through Steps 1, 2, and 3 Walk through three high-signal data science stories.
3. SLA & Detours 1:50 - 2:20 Violations & Geofencing Click "Next ▶" through Steps 4 and 5 Pitch the business model: automated compliance auditing.
4. Standalone Maps 2:20 - 2:45 TLV Speed & Bunching Click "🗺️ TLV Bus Speed Map" in header Show the massive, full-scale TLV telemetry dashboards.
5. The Appendix 2:45 - 3:00 Source Material & Wrap Close tab, scroll to bottom to show Source Material Prove 100% reproducibility and collaboration.

🗣️ Step-by-Step Script & Actions

0. The Hook & Civic Grounding (0:00 - 0:25)

  • Setup: Start on the main React page.
  • Action: Direct the audience's attention to the header, and click the blue "📰 Data Journalism Editorial" button. This opens Lihai's article in a new tab.
  • Spoken Script:

    "Good afternoon, judges. Today we are presenting the Open Bus Shared Infra Platform. Public transit open data is often treated as a raw, academic exercise—but we wanted to ground our research in real-world civic impact.

    We start with Lihai's Data Journalism Editorial: a long-form public narrative outlining what the open data actually shows about Israeli bus service. But Lihai's article isn't static—every chart and headline numbers is backed by a live, production-ready analytical engine we built together."

1. The Production Platform (0:25 - 1:00)

  • Action: Close Lihai's tab to return to the React dashboard. Point to the "Lines: 142" and "Operators: דן" filters. Then, click the orange "🖥️ Start Pitch Mode" button.
  • Spoken Script:

    "When we started this hackathon, we each had separate, messy Jupyter notebooks and drafts. Instead of presenting isolated slides, we consolidated all of our work into one unified, real-time dashboard carrying 17 registered analyses.

    When we change a single filter—like Line 142 of Dan—every single card, map, and heatmap synchronizes instantly. To walk you through our discoveries, we built Pitch Mode directly into our interface."

2. Narrative Case Studies: From Timetable to Poisson Decay (1:00 - 1:50)

  • Action: Click "Next ▶" on the Pitch Bar to go to Step 1 (Optimistic Timetable). The screen will automatically scroll and flash-highlight the Segment Reliability card.
  • Spoken Script:

    "Our first discovery is The Optimistic Timetable. In 'Where the timetable is optimistic', we compare GTFS scheduled times against actual GPS arrivals. We immediately see that schedulers are chronically optimistic about intermediate segments, leaving no margin for traffic.

  • Action: Click "Next ▶" to go to Step 2 (Rush Hour Breakdown). The screen scrolls and flashes the Heatmap. Hover over a dark red cell to show the new tooltip formatting.
  • Spoken Script:

    How do those delays form? In 'Which segments break down at rush hour', we continuously map travel-time ratios. We can see the exact segments that collapse during the morning rush hour. Hovering over a cell shows you not just the ratio, but the exact observed actual vs. planned minutes—and even seconds—lost.

  • Action: Click "Next ▶" to go to Step 3 (Regularity Decay). The screen scrolls and flashes Yuval's Poisson card.
  • Spoken Script:

    As these delayed buses travel downstream, headway spacing decays. In Yuval’s Poisson Arrival Regularity card, we map the headway Coefficient of Variation ($C_v$). Buses leave the origin perfectly spaced, but by the end of the route, $C_v$ approaches 1.0—indicating fully random, exponential Poisson spacing. The schedule is completely lost."

3. Business Model: Automated SLA Auditing (1:50 - 2:20)

  • Action: Click "Next ▶" to go to Step 4 (SLA Audits). The screen scrolls and flashes the Service Violations card.
  • Spoken Script:

    "This headway decay leads directly to severe bus bunching and contractual SLA infractions. In Israel, operators are heavily fined by the Ministry of Transport for early departures and cancellations.

    We built an Automated SLA Audit Tool that scans GPS pings on the fly. It filters out pre-departure boarding pings to prevent false alarms, and flags exact early departures, late terminal starts, and cancelled ghost rides. This is a ready-to-use contract enforcement platform for municipalities."

  • Action: Click "Next ▶" to go to Step 5 (Route Divergences). The screen scrolls and flashes the Divergence map.
  • Spoken Script:

    If a bus takes an unauthorized detour, our geofencing algorithm detects route divergence, plotting the exact physical streets where the driver strayed from the GTFS shape."

4. Standalone Tel Aviv Dashboards (2:20 - 2:45)

  • Action: Scroll to the top and click the blue "🗺️ TLV Bus Speed Map" button. This opens the speed map in a new tab.
  • Spoken Script:

    "Finally, to show the scale of our data engine, we integrated two full-scale interactive Tel Aviv dashboards. This is our Bus Speed Map, analyzing door-to-door speeds across every street in Tel Aviv for over 100,000 rides. It calculates exactly how many 'bus-minutes' are lost in gridlock, showing planners exactly where a bus lane is economically justified."

5. Open Source Collaboration & Wrap (2:45 - 3:00)

  • Action: Close the Speed Map tab. Scroll to the absolute bottom of the React dashboard to display the "Source Material" appendix. Expand one of the files (e.g. open_bus_poisson_analysis_all_in_one.ipynb) to show the code cell rendering.
  • Spoken Script:

    "To ensure 100% data transparency and peer reproducibility, we built a Source Material appendix at the bottom of the page. It connects server-side to fetch our original Jupyter notebooks, scripts, and static drafts—displaying our raw research side-by-side with our final interactive cards.

    We didn't just build a dashboard; we built a collaborative data infrastructure for public transit accountability. Thank you, and we welcome your questions!"

💡 Pro Presenter Tips for the Q&A Session:

  1. "Why do we see different sample sizes in the Heatmap?"
    • Answer: "Our data doesn't hide telemetry realities. Dropouts are due to poor cellular reception or cellular canyons. Furthermore, buses passing too quickly between 30-second SIRI pings skip the narrow geofences of intermediate stops, which naturally drops their sample sizes."
  2. "How do you resolve ambiguous line numbers?"
    • Answer: "We engineered a geographical dropdown filter 'City / route contains'. By choosing a city like 'תל אביב', our backend automatically resolves the correct line variant, preventing the uvicorn API from mixing up routes."
  3. "Can we scan more days?"
    • Answer: "Yes! To prevent heavy network loads during a live demo, we sample two weekdays by default, but we've built a 'Max days' option directly onto the bunching card so you can disable sampling and scan the entire date window on the fly."