Public space reports

Municipality of Heemskerk

Anyone can use the Fixi app or website to make a report to the municipality about public spaces (e.g. about litter, a defective lamppost or water nuisance). A dangerous traffic situation or nuisance can also be reported. Fixi is the municipality's notification system for resolving reports in public spaces.

Last change on 11th of February 2026, at 10:18 (CET) | Publication Standard 1.0

• Publication category: Other algorithms
• Impact assessment: Field not filled in.
• Status: In use

General information

Theme

Organisation and business operations

Begin date

09-2029

Contact information

[e-mail]

Link to publication website

https://www.heemskerk.nl/privacy

Responsible use

Goal and impact

If something needs to be fixed or cleaned up on the street or in a park, this can be reported to the municipality via the Fixi app, the municipality of Heemskerk's notification system. The notifier describes his report, the algorithm(1) recognises words from the description and suggests the main category of the report. The caller ultimately decides which main category fits the report best (e.g. 'nuisance and enforcement' or 'street furniture'). A professional group is linked to each main category so that the report goes to the right municipal handling team. A predetermined processing time is linked to each (sub-)category. Depending on the notifier's choice, the report is visible on the public map and thus visible to other residents (with the exception of nuisance and waste reports, which are always invisible). The algorithm(2) recognises personal data in the description and personal characteristics (license plates, faces) in photos and renders them unreadable.

Considerations

Previously, reporters had to choose which category their report best fit (e.g. 'nuisance' or 'street furniture'), so that the report went to the right department in the municipality. But the municipality is a complex organisation and the list of main categories is long. As a result, people did not always choose the right main category. This sometimes caused delays in processing reports. Therefore, we now use an algorithm(1) that recognises words, e.g. 'rubbish' and 'pavement'. Based on this, it suggests which main category best suits the report. The caller remains in control, however, and can still choose another main category if they disagree with the main category suggestion. Using the algorithm in Fixi makes the processing of reports about public space faster, more accurate and clearer. Reports are automatically recognised and classified, allowing the right department to get to work faster. As a result, the municipality can work more efficiently and reports are followed up better. Other ways of processing, such as fully manual categorisation of notifications, would take more time and capacity and lead to longer processing times. The deployment of this algorithm is therefore considered reasonable and justified.

To prevent any personal data in the description or photos from being visible to other residents, an algorithm(2) is deployed for blurring elements in photos and making data unreadable. Other ways of processing, such as manually checking all photos and texts, would take more time and capacity, with the added risk that reports made outside working hours would not be assessed and adjusted until the next working day. The deployment of this algorithm is therefore considered reasonable and justified.

Human intervention

Each report is still reviewed internally and assessed by the handler when the report is processed. Notifications that have arrived in the wrong main category are converted manually by adjusting the main category so that they go to the correct department.

Risk management

There are few risks in the algorithm (1). The algorithm has no impact on the decision, but suggests the right category and thus makes it easier to make the right choice. Thus, the notification is more quickly brought to the attention of the right department .

The algorithm (2) has no impact on a decision, but is intended to make any personal data unrecognisable in the description and on photos. Municipal employees see the original photos and dates. Residents see the censored photos and text, when the notification has been shared on the public map. The algorithm may not properly recognise a personal data, this will be acted upon on a situation-by-situation basis by the handler.

Operations

Data

The algorithm in the Fixi app processes data on reports in public spaces. This data is mainly about the type of report (such as broken lights or litter), the location and when the report was made. The information is used to automatically categorise reports and forward them to the right department or employee.

In addition, the submitter can choose to also provide personal data, such as name, address (not mandatory), phone number and/or e-mail address. These data are only used to contact you about the report if necessary.

It is up to the submitter whether these personal data are included: a notification can also be filed without sharing these data. Once a report is submitted, all personal data are automatically anonymised on the public map.

Technical design

Algorithm (1)

The Azure Language Studio Service (Multi class classification) is used for training and predicting notifications. When implementing the 'Automatic categorisation of notifications' module, a dataset of 250 descriptions is imported. These descriptions are linked to the categories. Every month, the descriptions and categories of handled calls are added to the training set, so that the category is predicted better and better.

Algorithm (2)

Number plate recognition is performed with an ALPR module that combines OpenCV's edge detection techniques with Pytesseract's LSTM-based OCR engine for accurate character extraction. Face detection is performed with the RetinaFace library, which uses a deep learning-based single-shot detection algorithm optimised for localising facial features. For text detection, we use EasyOCR, which uses the CRAFT algorithm for identifying text regions and CRNN for recognising text content. To ensure privacy, sensitive areas such as faces, number plates and specific text elements are blurred:

Gaussian convolution for image smoothing,

Colour space transformations for preprocessing,

and bitmap rendering for annotations such as rectangles and labels.

For text masking, we use a Named Entity Recognition (NER) module with the BERT algorithm, which provides contextual understanding for identifying sensitive textual entities. These entities are then masked using regular expression-based pattern matching algorithms.

External provider

Decos