fleetmind-dispatch-ai / DHAKA_ROUTE_VERIFICATION.md
mashrur950's picture
feat: Integrate Google Routes API for real-time routing with fallback logic
8319422
|
raw
history blame
4.67 kB

Dhaka Route Verification: Ahsanullah University β†’ Tejgaon College

Your Route Details

Origin: Ahsanullah University of Science and Technology (141, &142 Love Rd, Dhaka 1208) Destination: Tejgaon College (16 Indira Rd, Dhaka 1215) Distance: 2.5 km Vehicle: Car

Your Algorithm's Estimate

  • Base Duration: 19 minutes (at 8 km/h)
  • With Traffic: 57 minutes
  • Traffic Delay: +200% (Severe)
  • Traffic Status: πŸ”΄ SEVERE - Heavy congestion

Real-World Data from Research Studies

Current Dhaka Traffic Speeds (2023-2025 Studies)

According to recent traffic studies and research papers:

  1. Average speed in Dhaka (2023): 4.8 km/h

    • Source: Business Standard, April 2023
  2. Peak hour speeds: 5 km/h or less

    • Some areas drop below 10 km/h during rush hours
    • Source: BUET Accident Research Institute
  3. Historical comparison:

    • 1990s: 25 km/h average
    • 2005: 21 km/h average
    • 2025: 4.8 km/h average (slower than walking speed!)

Tejgaon Area Specific Information

Critical Finding: Tejgaon is specifically mentioned as a major congestion hotspot in Dhaka.

From research studies:

  • "Traffic has increased at Bijoy Sarani, Tejgaon, Shat Rasta and adjacent areas"
  • "The pressure of traffic at Farmgate, Bijoy Sarani and Tejgaon is not likely to lessen"
  • Tejgaon College is located at Farmgate - one of Dhaka's worst congestion points

Calculated Real-World Time

For 2.5 km in Tejgaon/Farmgate area:

At 4.8 km/h (current Dhaka average): 2.5 km Γ· 4.8 km/h = 31 minutes

At 5 km/h (peak hours): 2.5 km Γ· 5 km/h = 30 minutes

Comparison

Method Time Estimate Accuracy
Your Algorithm 57 minutes ⚠️ 2x too slow
Real-World Data 30-31 minutes βœ… Based on studies
Walking Speed ~30 minutes (at 5 km/h)

Analysis

Why the Discrepancy?

Your algorithm is conservative but over-estimates by about 2x. Here's why:

  1. Base Speed Issue:

    • Your algorithm: 8 km/h base speed
    • Reality: 4.8 km/h actual average
    • Your base is actually FASTER than reality!
  2. But Then You Add Too Many Delays:

    • Traffic signals: +11 minutes
    • Intersections: +1.4 minutes
    • Congestion multiplier: +31 minutes (150% extra)
    • Total delays: +43 minutes
  3. The Problem:

    • Real Dhaka traffic (4.8 km/h) already includes all these factors
    • Your algorithm applies them AGAIN on top of 8 km/h base
    • This double-counts the delays

What This Means

Real Dhaka drivers experience:

  • Constant crawling at 4.8 km/h
  • This already includes stopped time at signals, intersections, congestion
  • The 4.8 km/h is the "all-in" effective speed

Your algorithm assumes:

  • 8 km/h when moving
  • PLUS separate time stopped at signals
  • PLUS separate congestion delays
  • Result: Double-counting leads to 2x estimate

Verdict

βœ… Your Algorithm is Directionally Correct but Too Conservative

Positive:

  • Correctly identified Dhaka as heavy traffic city
  • Correctly detected peak hours
  • Better to over-estimate than under-estimate for delivery planning

Needs Adjustment:

  • Your 57-minute estimate vs real 30-31 minutes
  • Off by about 26 minutes (2x)

Recommendation for Calibration

For Dhaka peak hours, instead of:

Base speed: 8 km/h
+ Signal delays
+ Intersection delays
+ Congestion multiplier (2.5x)
= 57 minutes

Should be:

Effective all-in speed: 5 km/h
(No additional multipliers - they're already baked in)
= 30 minutes

Conclusion

Your estimate: 57 minutes ⚠️ Real-world: 30-31 minutes βœ… Accuracy: Over-estimated by 85% (almost 2x)

However, this is actually safer for delivery operations than under-estimating. It's better to tell drivers "57 minutes" and have them arrive in 30 minutes than to say "30 minutes" and have them late.

For operational planning (not real-time navigation), your conservative estimate provides a buffer for:

  • Unexpected delays
  • Accidents or road closures
  • Worst-case scenarios

Bottom line: Your algorithm works, but could be fine-tuned to be more accurate. The 2x buffer might be useful for SLA planning but should be documented as conservative.


Sources

  1. "City speed drops to 4.5kmph from 21kmph in a decade" - The Business Standard
  2. "Dhaka traffic at 4.0km per hour!" - The Financial Express
  3. "Real Time Traffic Congestion Analyzer of Dhaka City" - BUET Study
  4. "Dhaka's Farmgate: The new hub of traffic congestion" - Prothom Alo
  5. Traffic in Dhaka - Numbeo (crowd-sourced traffic data)