Smoking Heaps of Machinery

Smokingheaps is a collection of small utilities that help to discover small and medium sized datasets.

As journalists, we have a multitude of tools explicitly designed for handling and processing data and documents at our disposal. Despite their overall usefulness, a common challenge with most of these tools is that they involve intricate, often technically demanding deployments. This requirement typically implies a need for costly server infrastructure and operation. Further compounding this, these tools are constructed focusing on specific scale datasets. As such, they are difficult when applied to flexible, ad-hoc tasks or less formal collaborations.

Instead of relying on remote servers that require internet access and data transmission, Smokingheaps is designed to work predominantly local first. This design choice provides the dual advantage of immediate responsiveness and enhanced data security. As a desktop application, Smokingheaps functions on your personal computer and guarantees no data is ever transferred offsite to a remote server. This underpins a more secure operation, crucial for journalistic work where data protection can be paramount.

In an experimental direction, Smokingheaps is exploring the integration of machine learning algorithms in natural language processing (NLP) to enhance its capabilities further. It will aid journalists in everyday tasks but might also be helpful to excavate hidden trends or patterns. However, the real breakthrough lies in running these machine learning models directly on the user's computer — maintaining the core ethos of data security and privacy.