4 ways of using Artificial Intelligence to understand people in citiesDaoura

We are living a boom in people’s posts on social media and social interaction channels on the Internet, now fueled by the global health crisis that made millions of people around the world stay at home.

If they already used the internet to manifest on various topics before, now with the quarantine adopted globally, digital channels have become an even more relevant source to understand the experience and perception of people on social and urban issues.

But, how is it possible to understand that immense amount of information published all the time on the Internet?

The answer to that is: Artificial Intelligence. With the advancement in technology that allows machines to interpret texts written by us, human beings, it became much easier to extract relevant information from social media posts for a deeper understanding about what people say on the Internet.

Hereunder, Daoura highlights 4 practical examples of how companies and governments in four South American countries can understand the pulse of the cities and make smarter decisions, due to Artificial Intelligence.

1 - Measure the sentiment of the texts published by citizens

We express sentiments when we write. Those sentiments can be negative, positive or neutral. Thanks to Daoura's sentiment detection technology, digital manifestations about 10 cities in the 5 regions of the country are being analyzed under the sentiment perspective, so that it can be understood how their citizens manifest in general.

Distribution of sentiments detected in the manifestations of monitored Brazilian cities.

2 – Understand the concentration and distribution of the citizen’s opinions by city or region

Knowing the difference between how people think in different cities of a same country is important to understand the regional differences and take actions according to the needs and priorities of each local. That’s the case of Chile, country where Daoura monitors more than 14 cities to understand how people of different regions of the country think of the constitutional plebiscite, that will be voted on October.

Distribution of the concepts "apruebo" (approve) and "rechazo" (reject) associated with the new constitution found in the manifestations of monitored Chilean cities.

3 - Detect categories that summarize the texts according to the urban context

Classifying Internet manifestations into categories is important to understand on which topic people talk the most and how that topic varies with time. The Daoura Insights platform makes that categorization by urban dimensions (Health, Mobility, Education and Security, for example) and shows the concentrations and variations of those categories with time, to help cities of all sizes to understand where the main pains of citizens live. Some of Argentina’s major cities already count with this follow-up!

Frequency of manifestations by urban categories detected in demonstrations in cities of Argentina.

4 – Identify the most frequent words and expressions in manifestations

Knowing the main words and most frequent expressions in a set of manifestations allows us to understand which ones the citizens talk the most about in a city. When we combine this data with the predominant sentiment in the texts that contain a determined word, we can know if that topic is seen as good or bad by the population of a city. Cities from north to south of Peru are already tracked by the Daoura Insights Platform and identify the main topics and their respective sentiments.

Word cloud of the most relevant words and expressions and their sentiments found in the manifestations of the monitored Peruvian cities.