Summary  
This chapter covers data filtering and deduplication by applying reliability, timeliness, relevance, and uniqueness criteria to a collection.  

General domain of usage  
Data cleaning and quality assurance

Research may include outdated, unreliable, or repeated sources. Removing them keeps the project focused and improves the quality of the final report. Claude should remove sources that:

* do not identify an author or organisation;
* provide no evidence for their claims;
* are outdated for the research topic;
* repeat information from a stronger source;
* are not relevant to the research questions.

When several sources report the same information, Claude should keep the most reliable and original one.

Note

> ### Task
> 
> * Ask Claude to review the collected sources, remove weak or duplicate entries, and explain each decision.
> 
>   > **Review the sources in my workspace. Remove unreliable, outdated, irrelevant, or duplicate sources. Keep the strongest source for repeated information and briefly explain what you removed.**

Build a complete AI-powered research system using Claude Code. You will learn how to turn a broad topic into focused research questions, create a structured plan, find and evaluate reliable sources, and organise evidence in a reusable workspace. Through practical AI Lab chapters, you will create source libraries, compare conflicting information, store findings in a research database, and produce clear research briefs with accurate citations. By the end of the course, you will have a repeatable Claude Code workflow that can be used for future research projects.

Learn what makes a research system reliable and how to prepare a structured workspace before collecting information. You will define research questions, set the scope of the project, and create a clear research plan for Claude to follow.

Learn how to find relevant information and judge whether a source is trustworthy. You will remove weak, outdated, and duplicate sources before organising the strongest results into a structured source library.

Turn collected sources into clear and usable evidence. You will extract key findings, compare sources that disagree, and build a research database that keeps claims, evidence, and source details connected.

Use the collected evidence to develop supported conclusions and create a concise research brief. You will review claims and citations for accuracy before turning the complete process into a reusable Claude Code workflow.