Scientific Data Exploration

Argos

Scientific data exploration powered by Elasticsearch

Search and explore large datasets of organisms, viruses, bacteria, and related biological information through an interactive React interface.

Role

Software Engineer

Company

Dzango Tech Accelerator

Ecosystem

Embleema

Focus

Search · Data · React

Argos scientific data exploration interface

Context

Built within the Embleema ecosystem

Argos was a sub-project within the Embleema ecosystem. Embleema is a health-data and real-world-evidence platform that brings together complex healthcare and research data — including clinical, genomic, patient-reported, and other research datasets — for clinical and scientific use.

I worked on Argos during my time at Dzango. My contributions focused specifically on the Elasticsearch search infrastructure and the React frontend for biological data exploration. Embleema itself is a broader platform built by its own team.

Explore Embleema

Problem

Making large scientific datasets explorable

Scientific datasets can become difficult to navigate when users need to combine multiple dimensions of information simultaneously. Researchers required a way to locate relevant organisms and biological records using multiple criteria — not just a single identifier or a raw list.

The data covered organisms, viruses, bacteria, biological taxonomy, and related scientific metadata. Users — including doctors and scientists — needed to search and filter across attributes such as location, family, date, and biological classification to identify records relevant to their work.

A simple database lookup was not sufficient for this type of exploration. The system needed an efficient search and indexing strategy, and a frontend capable of presenting complex search results clearly.

Search Infrastructure

Elasticsearch indexing & search

Search pipeline

Scientific Dataset

Organisms · Viruses · Bacteria · Biological Taxonomy

Data Processing & Indexing

Transformation and preparation for search

Elasticsearch Index

Structured, queryable document store

Search & Filter Queries

Multi-attribute retrieval across large collections

React Application

Research-oriented exploration interface

Researchers & Scientists

Doctors · Scientists · Clinical researchers

What I worked on

My work on the Elasticsearch side of Argos centred on making large scientific datasets searchable and usable. This involved indexing biological records, designing the structure of Elasticsearch indexes, and implementing the search and filter queries that powered the frontend experience.

Indexing large datasets of organisms, viruses, and bacteria

Designing Elasticsearch index structure for scientific data

Implementing multi-attribute search queries

Supporting filtering by location, family, date, and classification

Structuring indexed data for efficient frontend consumption

Supporting retrieval performance across large record collections

Key principle

Data modelled for search is structured differently from data modelled for storage. A significant part of this work was understanding how to represent scientific records in a way that made them efficiently retrievable across multiple query dimensions.

Frontend

React research interface

The React frontend I worked on gave researchers and scientists a dedicated interface for exploring the biological dataset. Rather than exposing raw data, the interface was designed around a search-and-filter workflow that matched how researchers actually approach dataset exploration.

The interface was search-oriented, filter-heavy, and data-dense — designed for users who need to combine multiple criteria to locate relevant records efficiently. Users could search organisms and narrow results using location, biological family, date, classification, and other available metadata.

Interface characteristics

· Search-oriented — query first, browse second

· Multi-filter — location, family, date, classification

· Data-dense — structured results for researcher workflows

· Designed for scientific users, not general audiences

Research workflow

Query organisms by name, identifier, or attribute

Filter

Narrow results by location, family, date, or classification

Review results

Browse ranked, structured search output

Inspect record

View full biological metadata for a selected organism

Explore

Navigate related scientific information and taxonomy

Technical

Engineering challenges

Searching large datasets

Making large collections of scientific records searchable required designing an appropriate indexing strategy and query architecture. A simple database lookup was not sufficient for the type of multi-attribute exploration the project required.

Complex multi-dimensional filtering

Users needed to combine multiple independent dimensions simultaneously — location, biological family, date, and classification. This required careful query construction to produce correct and performant results.

Data-dense research interfaces

Scientific search interfaces can quickly become overwhelming when displaying dense, structured result sets. The React interface needed to balance information density with clarity, making complex data navigable rather than difficult.

Reflection

What I learned

Designing search experiences around actual user workflows rather than data structure
Elasticsearch indexing — how data is modelled for search differs from how it is modelled for storage
Constructing queries that support multi-dimensional filtering efficiently
Working with large, heterogeneous scientific datasets
Building data-dense React interfaces that remain usable under information load
Translating complex scientific records into interfaces that non-technical researchers can navigate

All projects View experience