Why Every Software Tester Should Learn Tableau in 2026

// Fail Flag
IF [Status] = "FAIL" THEN 1 ELSE 0 END

// Fail Rate (per test, independent of view filters except context)
{ FIXED [Test Name] : AVG([Fail Flag]) }

// Runs per Test
{ FIXED [Test Name] : COUNTD([Run Id]) }

// Flaky Candidate (thresholds are examples, agree with your team)
IF [Runs per Test] >= 20 AND [Fail Rate] > 0.05 AND [Fail Rate] < 0.95
THEN "Flaky?" ELSE "Stable" END

# Convert JUnit XML reports from CI into a CSV for Tableau
import csv, glob
import xml.etree.ElementTree as ET

with open("test_results.csv", "w", newline="") as out:
    w = csv.writer(out)
    w.writerow(["run_id", "test_name", "status", "duration_s"])
    for path in glob.glob("reports/run-*/junit.xml"):
        run_id = path.split("/")[1]
        for case in ET.parse(path).getroot().iter("testcase"):
            failed = case.find("failure") is not None or case.find("error") is not None
            w.writerow([run_id, f'{case.get("classname")}.{case.get("name")}',
                        "FAIL" if failed else "PASS", case.get("time")])

// p95 response time
PERCENTILE([Elapsed ms], 0.95)

// p50 response time
MEDIAN([Elapsed ms])

// Error Flag (JMeter "success" column is true/false)
IF [Success] = FALSE THEN 1 ELSE 0 END

// Error Rate
AVG([Error Flag])

import pandas as pd

# JMeter results (CSV/JTL) with columns: timeStamp, elapsed, label, success
df = pd.read_csv("results.jtl")
df["build"] = "2.4.1"                      # tag each run with its build
df = df.rename(columns={"elapsed": "elapsed_ms", "label": "endpoint"})
df[["build", "endpoint", "elapsed_ms", "success"]].to_csv(
    "perf_2.4.1.csv", index=False)
# Combine several builds' files in Tableau with a union.

Dashboard: Automation Health
Sheet 1: Bar — top 15 tests by fail rate (runs >= 20)
Sheet 2: Highlight table — test x run, pass/fail colour
Sheet 3: Line — overall pass rate by day
Sheet 4: Text — "3 tests to fix first" with reasons
Filters: suite, branch, date range
Publish to Tableau Public with a short description

import random, csv

random.seed(7)
tests = [f"test_{i:03d}" for i in range(60)]
flaky = set(random.sample(tests, 6))
with open("synthetic_ci.csv", "w", newline="") as f:
    w = csv.writer(f)
    w.writerow(["run_id", "test_name", "status"])
    for run in range(1, 101):
        for t in tests:
            p_fail = 0.25 if t in flaky else 0.01
            w.writerow([run, t, "FAIL" if random.random() < p_fail else "PASS"])

Week 1: Connect to CSV, dimensions vs measures,
        basic bar and line charts
Week 2: Calculated fields, IF statements, aggregation
Week 3: LOD expressions (FIXED), percentiles,
        parameters for thresholds
Week 4: Build the Automation Health dashboard and
        publish it on Tableau Public

// Let viewers change the flaky threshold
// Create parameter [Flaky Threshold] (float, default 0.05)
IF [Runs per Test] >= 20
   AND [Fail Rate] > [Flaky Threshold]
   AND [Fail Rate] < 1 - [Flaky Threshold]
THEN "Flaky?" ELSE "Stable" END

- Built a Tableau dashboard analysing synthetic CI
  results to rank flaky tests by fail rate (LOD
  calculations) and compare performance percentiles
  across builds; published on Tableau Public