How Entity Reputation Tracking Works
Our Entity Reputation Timeline tracks how any public figure — politicians, business leaders, celebrities — is portrayed across Indian media over time. We identify every mention of an entity, analyze the sentiment of the surrounding coverage, and plot it on a timeline so you can see shifts in media treatment.
Sentiment Timeline
Each data point on the timeline represents the average sentiment score (0-100) for coverage of that entity on a given day. Below 40 indicates negative coverage, 40-60 is neutral, above 60 is positive. Watch how major events trigger sentiment shifts.
Source-Level Breakdown
See how individual sources cover the same person differently. One outlet may consistently portray a politician positively while another focuses on criticism. This source-level breakdown reveals which outlets are sympathetic vs adversarial toward specific figures.
Methodology: How Entity Reputation Is Tracked
Entity reputation tracking works by identifying every mention of a person or organisation across all articles, analysing the sentiment of the surrounding text, and aggregating those sentiment scores over time. The result is a day-by-day timeline of how the media portrays any public figure, computed from real coverage data rather than editorial opinion.
Entity Detection
Our named entity recognition system identifies mentions of people, organisations, political parties, and other entities within article text. The system disambiguates entities so that "Modi," "PM Modi," and "Narendra Modi" are all recognised as the same person. Entity mentions are linked to the surrounding sentences, which are then scored for sentiment to determine whether the coverage is positive, negative, or neutral.
Sentiment Scoring
Sentiment is scored on a 0-100 scale for each entity mention. Values below 40 indicate negative coverage, 40-60 is neutral, and above 60 is positive. The system analyses the tone of sentences containing the entity mention, not just whether the entity appears in a positive or negative story overall. This entity-level sentiment is more precise than article-level sentiment because a single article may cover multiple entities with different tones.
Timeline Aggregation
Daily data points on the timeline represent the average sentiment score across all mentions of the entity on that date. Significant events are automatically detected when the daily sentiment deviates substantially from the rolling average. These spikes and drops often correspond to real-world events such as policy announcements, scandals, or election results that shift media treatment of the entity.
How to Interpret the Results
Reading the Timeline Chart
The chart has two visual layers. The solid pink line tracks sentiment (0-100), plotted against the left axis. The dashed purple line tracks political lean of coverage, plotted against the right axis. A horizontal dashed line at 50 marks the neutral sentiment baseline. Yellow dots highlight significant events where sentiment shifted sharply. By examining where sentiment dips or spikes, you can correlate media treatment with real-world events and assess whether coverage shifts are proportionate.
Source-Level Breakdown
Below the timeline, the Coverage by Source section shows how individual news outlets cover the same entity. Each source displays its average sentiment score and article count. This reveals which outlets are sympathetic versus adversarial toward specific figures. A politician might receive overwhelmingly positive coverage from one outlet and negative coverage from another. Comparing source-level sentiment helps you identify outlet-specific biases and build a more complete picture of media treatment.
Key Metrics Explained
The metrics panel at the top summarises four values. Total Articles shows the volume of coverage. Current Sentiment shows the latest day's average score. Trend shows the direction of sentiment change over the selected period (positive means improving coverage, negative means worsening). Volatility measures how much sentiment fluctuates day to day, with higher values indicating more unpredictable coverage patterns that may reflect event-driven reporting or editorial inconsistency.
Frequently Asked Questions
How do you track media coverage of politicians?
We identify mentions of entities (people, organizations) across all articles and analyze the sentiment of each mention. The timeline shows how positive or negative coverage changes over time, helping identify media bias patterns.
Is media biased against Modi or Rahul Gandhi?
Rather than making claims, we show you the data. Search for any politician and see their sentiment score over time across all sources. Compare coverage to form your own conclusions about potential bias.
What does the sentiment score mean?
Sentiment ranges from 0-100: below 40 is negative coverage, 40-60 is neutral, above 60 is positive. We analyze the tone of sentences mentioning the entity, not just whether they appear in positive or negative stories.