nxtl-esim/mobile-data-usage-estimates

Mobile Data Usage Estimates

The activity-level MB/hour assumptions and daily presets used by the open-source trip-data-budget and team-data-planner tools — published so other apps and agents can reuse the same transparent numbers.

Every figure is an estimate, not a measurement.

Tables

activities (8 rows)

activity_id	activity	MB / hour
maps	Maps & navigation	8
msg	Messaging & email	10
web	Web browsing	60
music	Music streaming	70
social	Social media (with video)	150
call	Video calls	500
video	Video streaming (SD)	700
hot	Laptop via hotspot	300

presets (4 rows)

Hours/day mixes that the tools ship with, plus the daily total after a 1.1× buffer for background updates and sync:

preset	est. MB / day (with buffer)
light	59.4
normal	349.8
heavy	1718.2
remote	2219.8

Streaming assumes standard-definition video; HD can use several times more. Real use varies with app settings and network conditions.

Example uses

• In-app data budget calculator: reuse the same MB/hour table so your UX matches the public tool.
• AI agent: when a user says "I'll navigate all day and do two hours of calls", multiply the rows instead of inventing rates.
• Education / blog: cite a small, versioned table rather than hand-waving "about a gig a day".

Quick load

from datasets import load_dataset
acts = load_dataset("nxtl/mobile-data-usage-estimates", "activities", split="train")
presets = load_dataset("nxtl/mobile-data-usage-estimates", "presets", split="train")
print({r["activity_id"]: r["mb_per_hour"] for r in acts})

import pandas as pd
acts = pd.read_csv("activity_mb_per_hour.csv").set_index("activity_id")["mb_per_hour"]
hours = {"maps": 1, "msg": 3, "web": 1, "music": 1, "social": 1}  # "normal" mix
mb = sum(acts[k] * h for k, h in hours.items()) * 1.1
print(round(mb, 1), "MB/day estimate")

Methodology

Copied from the constants embedded in trip-data-budget (ACTS, PRE, 10% buffer). No separate lab measurement was performed for this release. Where a figure is used in product copy it is labelled as an estimate.

Limitations

• Not a substitute for measuring your own apps on your own network.
• No per-app breakdown (Instagram vs TikTok, Zoom vs Meet, etc.).
• Video is SD-only in the table; there is no HD/4K row.
• Presets are starting points, not personas backed by survey data.

Licence and credit

CC BY 4.0. Attribute "NXTL mobile data usage estimates".

Live tool: trip-data-budget · sibling planner: team-data-planner

Citation

@misc{nxtl_mobile_data_usage_estimates_2026,
  title        = {Mobile Data Usage Estimates},
  author       = {NXTL},
  year         = {2026},
  howpublished = {Hugging Face dataset},
  note         = {Assumptions from trip-data-budget / team-data-planner}
}