Hadith Network

Project Overview

Narrative Networks refer to a class of social networks which are formed by creating edges between people who have narrated stories, reports, incidences etc from one another. These can be historical or contemporary in nature. The graphs used to represent these networks are directed graphs where the incident node corresponds to the person from whom the narration was originally taken. Datasets related to narrative networks have not been traditionally available to the research community and thus progress in this area has been limited. This project seeks to address this issue by not only analyzing historical narrative networks but also making them available to the research community.

I am in the process of reconstructing historical social networks of narrators spanning over the course of a hundred years from an Arabic hadith books, from the 8th-10th century of Islam. A hadith refers to the sayings of Prophet Muhammad. Hundreds of thousands of such sayings were recorded in the collection of hadiths and each narration from Prophet Muhammad is recorded as a chain of narration from the person who wrote them down to the person who had heard the narration from Prophet Muhammad himself. The first text that we are analyzing is one of the most famous and authentic collection of hadiths called Sahih Bukhari as a test case for the analysis of narrative networks.

Project Goals

The goals of this project are two fold: Jump start the Science of Narrative Networks as a viable field of Social Network Analysis. While the field of social networks has been around since the 1920s, narrative networks have not been analyzed to any extent to date mainly because of unavailability of datasets. This project seeks to remedy this issue by reconstructing these networks from early Islamic texts and making them available to the scholarly community. The networks which are formed by the chain of narrations in the hadith literature constitute unique types of social information network which has not really been studied from a graph theoretic point of view in the scientific literature before. These narrative networks have similarities to pedigree networks as well as citation networks but are fundamentally different from both types of such networks. Some of the problems that can studied from a network analysis and data mining perspective are as follows: History and Evolution of Communities of scholars, Stylometrics of Hadith networks, Properties of Hadiths Narrative Networks, Authorities and their Impact Over Time, Relationship Prediction and Missing Link Prediction, Topic Models of Hadith Literature etc.

Secondly, the project can also be useful to the scholarly community studying the history of early Islam, especially the hadith corpus by providing a new set of tools from the network analysis and computer science domain. The canonical Sunni collection of hadith (collections of sayings of Prophet Muhammad) consists of six main collections (saha sitta) of hadiths and multitude of other texts. The collection of these hadiths was a systematic effort spread of many decades. Additionally the field of biography of narrators (Isma'ur'rijal) in Muslim scholarship arose as a way to determine the trustworthiness of the people who were part of the chains of narrations from the Prophet Muhammad to the person who was collecting the hadiths. The main way to authenticate the hadiths was to list and determine the trustworthiness of all the people who narrated the hadith from Prophet Muhammad to the person who recorded the hadith. As a result many of the canonical collections of hadiths (Sahih Bukhari, Sahih Muslim, Behaqi etc) can be used to construct hundreds of thousands of such chains. Analysis of such chains of narrations had been impossible in the past because of the time and resources involved in the endeavor and limitations in human information processing. With the advent of computers and advances in data mining as well as developments in the information flow networks it is now possible to overcome these limitations. Thus the field of data mining, analysis of social and information networks, and computational stylometry (use of computational techniques to study style of writing) offers new possibilities in the analysis of hadith literature.


I have submitted a couple of manuscripts which are under review currently. I will make the datasets available as soon as they get accepted. Since this is a massive undertaking I will be releasing the datasets at regular intervals.


Since hadith networks can be represented as graphs, visualizations of these graphs can be used to illustrate important properties of the networks as well as insights regarding the various nodes that constitute these networks. I am also creating a repository of such visualizations which would also be helpful in communicating some of the results to not only the scholarly community but also to the general public as well. Click on any of the images below to go to the visualization page

Wudu Network
Khauf Network
Muwaq Network
Jummah Network
Jummah Network
Jummah Network
Wudu Network
Wudu Network
Network visualization of Kitab Jummah and Kitab Wudu from Sahih Bukhari.


For more information regarding this project and how you can help the project, you can contact me (Muhammad Aurangzeb Ahmad) at mahmad@cs.umn.edu More information about my research and publications in other areas is available at the Publications Page.



I would like to acknowledge the following people for their help in transcribing the chain of hadiths: Tareg Banega, Mohamed Elbadry and Omar Mehana.