Water sensor data processing and analysis

Accelerating environmental monitoring with AI-driven analysis of water flow and rainfall sensor data

Scroll down to read more

Project highlights

Industry: Architecture, Engineering, and Construction

Client services: AI Business Transformation

Started in 2025

Location: Detroit, MI, USA

Team size: 7 members

Duration: 6+ months

About the client

The client is a civil and environmental engineering firm that supports sewer and stormwater infrastructure projects. Its teams work with large volumes of time-series data collected from flow meters, rainfall gauges, and related environmental monitoring sensors to support calibration, QA/QC, and engineering analysis. Based on the project documents, the client needed a more structured and scalable way to ingest, validate, edit, and export sewer flow-monitoring data used in calibration projects.

As monitoring programs expanded across sites and sensors, the client needed to move beyond fragmented spreadsheets and ad hoc scripts toward a unified digital platform that could support both current engineering workflows and future AI-driven analytics.

Business challenges

A key part of the client’s work involves handling large datasets generated by water flow and rainfall sensors deployed across sewer-system monitoring projects. These datasets must be ingested, validated, corrected, reviewed, and prepared for downstream modeling and reporting.

However, the process was largely manual and fragmented.

Typical workflows required engineers to:

import raw meter datasets from files and spreadsheets,
validate file structure and measurement consistency,
identify anomalies, outliers, and missing values,
compare raw and edited data across versions,
analyze rainfall and flow relationships,
prepare accepted datasets for export and calibration use.

Processing this information manually created several challenges:

significant engineering time spent on repetitive data preparation,
QA/QC turnaround that often took 2–3 days per site or batch,
inconsistent practices across teams and sites,
risk of rework, missed anomalies, and costly field revisits,
difficulty scaling workflows as sensor networks and data volumes grew.

The client wanted to modernize this workflow with AI and automation while preserving the engineering rigor required for sewer calibration and wet-weather analysis.

Goals set to Achievion

Achievion was engaged to design and develop an AI-powered platform that would give engineers a unified workspace for handling water and environmental sensor data.

Key objectives included:

ingest raw meter datasets from CSV and Excel files,
validate data structure and automate schema mapping, including LLM-based mapping recommendations,
identify outliers and apply rule-based quality checks,
support sensor data versioning, comparison, and updates,
provide rainfall, runoff, and hydrograph analysis tools,
create a scalable platform foundation for future advanced analytics and AI workflows.

The ultimate goal was to reduce manual engineering effort, improve data quality, and create a scalable digital foundation for intelligent infrastructure monitoring.

Solution

Achievion designed and implemented an AI-enabled water sensor data platform that automates the ingestion, validation, analysis, and review of time-series data from flow meters, rainfall sensors, and related monitoring devices.

The platform combines several intelligent capabilities into a single workflow that processes incoming environmental datasets and converts them into structured, reviewable engineering information. The platform designed as extensible system, enabling rapid delivery while creating a foundation for AI-driven features.

Key features of the product in detail

AI-assisted sensor data ingestion

The first stage of the platform focuses on bringing raw water and rainfall sensor data into a unified system.

Achievion implemented ingestion pipelines that support CSV and Excel datasets, validate file structure, and associate incoming data with the correct site and sensor using metadata or filename patterns. The platform also supports ingestion history, version comparison, re-upload, and standardized templates for mapping source data into the internal schema.

A notable feature is AI-based schema mapping recommendation, which helps accelerate onboarding of varied sensor file formats into a consistent data model.

Intelligent data quality and anomaly detection

Once data is ingested, the system applies automated quality controls to identify problematic readings and improve trust in downstream analysis.

The platform includes:

rule-based checks for time-series validation,
outlier detection,
anomaly flagging,
support for detecting seasonality patterns,
workflows for updating existing sensor data.

These features reduce the need for engineers to manually inspect every row of sensor output and instead focus attention on flagged issues and exceptions.

Advanced water-system analytics

We also implemented specialized analytics to support engineering use cases around sewer and stormwater monitoring.

The solution supports:

comparison of raw versus edited datasets,
hydrographs,
flow vs rainfall analysis,
rainfall vs runoff analysis,
event tables with total rainfall, peak intensity, peak flow, and total volume,
event categorization for scenarios such as small rainfall, winter conditions, data issues, and outlier responses.

These capabilities make the platform directly useful for hydrological interpretation and calibration workflows, not just generic data handling.

Visualization and engineer review workflow

To make insights actionable, Achievion delivered interactive visualization features that help engineers review and validate sensor behavior.

The platform supports:

multi-series charting,
zoom and pan functionality,
direct visual comparison of time series,
time-shift adjustments within visualizations,
documentation of engineer annotations and comments.

This gives teams a unified interface for both automation and expert review.

Scalable AI-ready architecture

The platform was designed as a scalable architecture with controlled environments, reusable pipeline components, time-series and metadata storage, and a low-code / no-code workflow capabilities for engineers.

That architecture creates room for future AI/ML enhancements such as:

improved anomaly detection models,
ML-based data regeneration,
predictive monitoring,
intelligent sensor health assessment,
deeper pattern analysis across water infrastructure datasets.

Business outcome

Achievion delivered an AI-enabled platform that transformed fragmented sensor-data workflows into a more scalable and efficient engineering process.

Key outcomes:

QA/QC turnaround reduced from 2–3 days to 2–4 hours,
usable data rates improved from 85–90% to 95–98%,
manual correction reduced from 50–70% of rows to 10–20%,
report delivery timeliness improved to 95–100%,
engineers were able to focus more on higher-value analysis instead of repetitive data cleanup.

The solution gave the client a scalable digital foundation for managing water sensor data, improving engineering productivity, and preparing for AI-driven monitoring and analytics.

Timeline

1 Month

Design Phase

Analyzed existing sensor data workflows
Assessed data quality issues
Defined system requirements, KPIs, and success metrics
Designed AI/ML approach for anomaly detection and data processing

1.5 Months

Prototype Development

Built initial data ingestion and preprocessing pipelines
Developed ML models for detecting anomalies (gaps, drift, outliers)
Created a prototype UI to visualize sensor data and AI insights
Validated AI approach with real-world datasets and stakeholder feedback

4 months

MVP Development

Developed core platform modules: ingestion, QA/QC, analytics, visualization
Implemented AI-assisted schema mapping and automated workflows
Delivered rainfall, flow, and event analysis tools
Deployed cloud-based application and conducted user demos for validation

Team

Product Manager

Solutions Architect

Data Scientist

UI Designer

Software Developer

MLOps Engineer

QA Engineer

Get in touch to learn how our AI powered solutions
can solve your business problem.