What I Learned from Analysing 80,000 Hourly Benzene Measurements

For the February 2024 Senate hearing, PeaceLink asked me to examine all the hourly benzene measurements from the Via Machiavelli monitoring station in Taranto across 2013–2023.

The dataset contained more than 80,000 observations.

That number sounds substantial, but volume was not the difficult part. The difficult part was deciding what could be claimed from the measurements, preserving the hourly events that annual summaries hide, and making the result clear enough to enter a public and institutional discussion without making it sound more certain than it was.

The analysis was later presented by Alessandro Marescotti during an Italian Senate hearing on 6 February 2024. The official slides credit my data analysis and processing with Omniscope.

What follows is not only the result. It is what the work taught me about public evidence.

Hourly grain changes the question

Benzene regulation and public reporting often focus on annual averages. Those are essential for assessing long-term ambient concentrations and legal compliance.

An annual mean cannot describe the shape of every hour inside it.

Two years can have the same annual average while one contains a relatively stable distribution and the other contains short, extreme episodes. People living near a monitoring station experience both patterns, but the analytical questions are different:

• What was the station’s annual mean?
• How often did unusually high hourly episodes occur?
• Were they concentrated in particular months or times?
• Did the distribution change?
• Were measurements missing during events of interest?

Aggregating too early destroys the evidence needed for the second group.

The first lesson was simple: retain the finest trustworthy grain, then derive the summaries. Do not keep only the summary and hope it answers every future question.

A threshold needs a name and a meaning

For the hourly analysis, PeaceLink used 27 µg/m³ as a reference.

That value comes from the California Office of Environmental Health Hazard Assessment’s acute Reference Exposure Level for benzene. It describes a one-hour airborne concentration not anticipated to cause adverse non-cancer effects for that exposure duration.

The wording matters:

• it is a health-based acute reference level;
• it is designed for infrequent one-hour exposure;
• it is not an Italian or EU ambient legal limit;
• exceeding it is not automatic proof of individual harm;
• exceeding it is not, by itself, proof of a legal breach or source.

The EU ambient benzene limit under Directive 2008/50/EC is expressed as an annual mean. The recast Directive (EU) 2024/2881 sets a tighter annual value to be attained by 2030.

An hourly 27 µg/m³ event and an annual 5 or 3.4 µg/m³ limit cannot be compared as if they were interchangeable thresholds. Different averaging periods answer different health and legal questions.

Whenever an alert or chart draws a line, the line should carry its source, duration and status.

What the February 2024 snapshot showed

In the dataset snapshot analysed for the Senate hearing, the Via Machiavelli series contained 63 hourly observations above 27 µg/m³.

Their yearly distribution was:

Year	Hourly readings above 27 µg/m³
2013	4
2014–2017	0
2018	1
2019	0
2020	2
2021	7
2022	17
2023	32
Total	63

These counts appear on page 5 of PeaceLink’s Senate submission.

The striking finding was the concentration in the latest year of that snapshot: 32 in 2023, compared with 31 across the previous ten years combined.

That justified attention. It did not identify a cause by itself.

Time, wind direction, other stations, industrial operations and validated source records all become relevant to a causal investigation. A single monitoring series can show when and where a pattern deserves examination. It cannot close every explanatory question.

A later official report gave a different count

There is an important complication.

The later final ARPA Puglia benzene report for 2023 records 47 Via Machiavelli observations above 27 µg/m³, not 32.

The exact ARPA Puglia 2023 benzene report contains that later official count.

The February hearing used an earlier dataset snapshot. I have not found a public reconciliation that explains the difference. It may involve validation, completion or revision of the source series, but the reason remains unresolved and should not be guessed away.

This changed one of my strongest lessons:

A result is not fully described by the query. It also needs the source snapshot and its date.

Anyone citing the 63/32 result should call it what it is: the dataset snapshot analysed and filed for the February 2024 hearing. The final 2023 ARPA report is a later version and should be cited for its final annual count.

The discrepancy does not make the earlier analysis worthless. Hiding it would.

Missingness is evidence too

Public environmental datasets contain gaps, station outages, delayed validation and revised records.

These must remain visible.

For each period I want to know:

• how many hours should exist;
• how many measurements are present;
• whether missingness is isolated or continuous;
• whether values are provisional or validated;
• whether timestamps and daylight-saving changes were handled consistently;
• whether duplicate observations exist;
• whether units changed;
• when the source file was retrieved.

A line chart that connects across a three-day gap can visually imply continuous monitoring. An annual summary can hide that the station was absent during an event somebody wants to investigate.

Absence does not prove why data is missing. It does affect what the available data can support.

A peak and a trend can both be true

Later analysis across 2023–2025 showed a reduction in the number and average of high daily benzene readings at several Taranto stations, particularly around Tamburi.

At the same time, later hourly episodes still exceeded the 27 µg/m³ acute reference.

These findings are not contradictory.

Longer-term improvement can coexist with individual events that deserve attention. Public communication becomes distorted when it insists on only one of those truths.

The distinction between hourly and daily measures belongs in the caption, not in a footnote nobody reads.

Evidence has to travel

The analysis moved through several forms:

1. source monitoring records;
2. a prepared and checked hourly dataset;
3. an Omniscope workflow and report;
4. charts and threshold counts;
5. PeaceLink’s public explanation;
6. the Senate hearing slides.

At every hand-off, context could have been lost.

The most valuable properties were:

• the raw observations remained available;
• the distribution by year was visible;
• the threshold was named;
• the station and period were explicit;
• the method could be rerun;
• later data could be compared;
• uncertainty about causation remained stated.

This is what turns a chart into public evidence rather than campaign decoration.

What I would preserve in the next analysis

For any long-running environmental series, I would now keep:

• the original downloaded files;
• retrieval time and source URL;
• file checksum where practical;
• validation or provisional status;
• normalised UTC and local timestamps;
• unit and averaging period;
• missingness report;
• transformation version;
• threshold definition and source;
• result snapshot used for each publication;
• links from every public claim back to the interactive evidence.

I would also reconcile later official revisions explicitly rather than silently updating an old number.

The number is not the conclusion

“More than 80,000 measurements” communicates scale.

“63 above the reference” communicates a finding from one snapshot.

Neither is the complete conclusion.

The complete work includes grain, coverage, threshold meaning, version, comparison and the limits of attribution. It includes the later official count that does not match the earlier snapshot. It includes the fact that trends can improve while significant episodes continue.

That may be less satisfying than one permanent headline.

It is much closer to evidence.