Skip to content
Tutorials Center
My learning

Topics

Lesson 26 of 36 ยท Timeline and Artifact Analysis

Timeline Analysis Basics

Learn why timelines are central to investigations: event types, time zones and clock drift, normalizing to UTC, merging sources, and pivoting around a key moment.

Evidence rarely speaks in a single voice. A login in one log, a file created on a disk, a DNS lookup in a network record and a message on a phone are each small facts. A timeline puts them in order so that cause and effect become visible: the phishing email arrived, the attachment opened, a process started, a connection went out. Timeline analysis is how separate artifacts become a story that a reader can follow and challenge.

What goes on a timeline

Every entry should have at least four parts:

Part Example
When (in UTC) 2026-11-12T14:03:11Z
Source web server access log, $MFT, CloudTrail
What POST /login 200, file created, key added
Where / who host name, user, address

Events come from many places: file system timestamps, logs, browser history, registry and system artifacts, email, network records and cloud audit trails (Windows artifacts covers many of these). Keep the raw source and original timestamp next to your normalized one, so anyone can check your conversion.

The hardest part: time itself

Most timeline mistakes are about time, not about the events.

  • Time zones. Some systems log in UTC, some in local time, some without any zone. Convert everything to UTC and note the original zone.
  • Daylight saving. Local times around a clock change can be ambiguous (the same hour occurs twice).
  • Clock drift. A machine's clock may be minutes off. If you can compare an event with a trusted source (for example a server log showing the same request), compute the offset and record it.
  • Different precision and epochs. Seconds vs milliseconds vs 100 ns ticks; Unix 1970, Windows FILETIME 1601, Apple 2001 (epoch conversion).
  • Meaning of the timestamp. "Modified" is not "created", and a log time is when the entry was written, not necessarily when the action started.

An investigator can change the system clock of a copy by accident, or misread a zone, and silently shift evidence by hours. State your time assumptions in the report and double check at least one event against an independent source.

Normalize and merge

The workflow is simple: parse each source, convert to UTC, tag it with its source, and sort.

Python Normalize events from three sources

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Sorted, the story reads: an upload to the web server, a new executable on the workstation 29 seconds later, and an outbound connection after that. Without the +1 conversion and the drift correction, the order would be wrong.

Pivot around a moment

Large timelines hold millions of rows, so you rarely read them top to bottom. Pick a pivot (the first alert, a known bad file) and look at everything within a window before and after it.

Python Window the timeline around a pivot

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The events at -29 s through +122 s form a cluster. The quiet entries at 14:00 and 14:40 fall outside the window and stay out of the way.

Reading a timeline well

  1. Look for clusters and gaps. A burst of activity is an incident; a gap in a normally busy log may be deletion.
  2. Ask what is missing. An executable appears but no download is recorded; where did it come from?
  3. Separate observation from inference. "File created at 14:03:40" is a fact; "created by the upload" is a hypothesis to support with other evidence.
  4. Test alternative explanations. Could a legitimate update explain the same sequence?
  5. Keep it reproducible. Save the script or tool commands used to build it.
flowchart LR
  A["Upload to web server"] --> B["Executable created"]
  B --> C["Persistence added"]
  C --> D["Outbound connection"]

A hypothesis diagram built from the timeline: each arrow needs supporting evidence.

Tools you will meet

Real investigations use tools that parse many formats into one timeline, such as Plaso/log2timeline and Timesketch, or spreadsheet-like viewers; see super timelines. Whatever the tool, the principles above decide whether the result is trustworthy.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. What should you do first when merging events from several sources?

    A common time base is required before sorting makes sense.
  2. 2. A firewall clock is known to run 90 seconds fast. How do you treat its events?

    Correct for known drift and document it so the reader can verify.
  3. 3. Why use a pivot window?

    Focusing on events around a key moment makes large timelines tractable.
  4. 4. Which statement is an observation rather than an inference?

    A recorded timestamp is an observation; the rest are hypotheses needing support.

Summary

  • A timeline orders events from many sources to reveal cause and effect.
  • Normalize to UTC and handle zones, daylight saving, drift and epochs.
  • Keep the original value, source and conversion for every row.
  • Pivot around a key event and look for clusters, gaps and missing links.
  • Separate facts from hypotheses.

Next: File System Timestamps and mactime