Data mining is one of the most used techniques to discover patterns in large datasets. The insights gained using data mining can be a valuable information for data-driven approaches in marketing as they provide an increase in promotion effect and customer loyalty. However, companies often consider the outcome of data mining as the ending point and do not use it for further investigations and generate an added value for improving decision making processes. The goal of this thesis is to demonstrate that it is possible to integrate the newly discovered patterns and customer segments with wide analyses of the socioeconomic situation of the area of interest, offering companies an useful tool to be consulted as assistance information for managers, according to the data-driven paradigm. The main advantage of the approach developed in this thesis is that it exploits free and publicly accessible datasets from national databases and open data sources to build a tool that can support retail managers in all marketing decisions, from locating the perfect spot for a new store to advertisement placement. To this extend, datasets on the Italian population are considered, and clustering algorithms are used to analyze the demographics of population, allowing to identify geo-referenced similarities. More in-depth analyses of the population distribution, positioning of stores, house prices and incomes distribution of Milan are also carried out, along with an analysis of the public transport of Lombardy. Analyzing the movements of people towards areas where stores are located allows to gain further insights: by integrating the public datasets of population and movements with real data on customers from stores it is possible to estimate the age group, reason for travel, mean of transport and origin of people entering a given store. It follows that a more precise identification of customers can result into specialized advertisement and/or a better located one. The results of these analyses have been made available in a website. All these insights can drive managers' decisions to a more data-driven and focused approach to marketing, as marketing efforts can be concentrated not only on specific customer segments generated by data mining tools but also on specific geographic areas.
Il data mining è una delle tecniche più utilizzate per scoprire relazioni insite in grandi quantità di dati e le informazioni acquisite possono rappresentare una parte preziosa per un approccio al marketing basato sui dati, con lo sviluppo di campagne promozionali mirate e una miglior fidelizzazione dei clienti. Tuttavia, le aziende spesso considerano queste informazioni come il punto finale dell'analisi e non le utilizzano in ulteriori indagini atte a migliorare i processi decisionali. L'obiettivo di questa tesi è dimostrare che è possibile integrare tali informazioni con analisi della situazione socioeconomica dell'area di interesse, offrendo alle aziende informazioni e suggerimenti tramite uno strumento online. Il principale vantaggio dell'approccio sviluppato in questa tesi è che sfrutta dataset gratuiti e pubblicamente accessibili da banche dati nazionali e fonti di dati aperte per costruire uno strumento in grado di supportare i gestori dei negozi in tutte le decisioni di marketing, dall'individuazione del miglior punto per un nuovo negozio al posizionamento pubblicitario. A tal fine, sono presi in considerazione dati sulla popolazione italiana e analizzati tramite algoritmi di clustering, consentendo di identificare somiglianze georeferenziate. Vengono inoltre effettuate analisi più approfondite sulla distribuzione della popolazione, sul posizionamento dei negozi, sulla distribuzione dei prezzi delle unità immobiliari e dei redditi di Milano, nonchè un'analisi del trasporto pubblico della Lombardia. Analizzare gli spostamenti delle persone verso aree specifiche permette di ottenere ulteriori approfondimenti: integrando dataset pubblici su popolazione e movimenti con dati reali di clienti è possibile stimare la loro fascia di età, il motivo del viaggio, il mezzo di trasporto e provenienza. Ne consegue che un'identificazione più precisa dei clienti può tradursi in pubblicità specializzata e/o meglio locata. I risultati di queste analisi sono stati resi disponibili in un sito web. Tutte queste informazioni possono guidare le decisioni dei manager verso un approccio al marketing più focalizzato e basato sui dati, poiché gli sforzi di marketing potranno essere concentrati non solo su segmenti di clienti specifici (generati da strumenti di data mining), ma anche su specifiche aree geografiche.
Exploiting openly available datasets for population clustering with demographic information and mobility data
BELLINI, SIMONE
2021/2022
Abstract
Data mining is one of the most used techniques to discover patterns in large datasets. The insights gained using data mining can be a valuable information for data-driven approaches in marketing as they provide an increase in promotion effect and customer loyalty. However, companies often consider the outcome of data mining as the ending point and do not use it for further investigations and generate an added value for improving decision making processes. The goal of this thesis is to demonstrate that it is possible to integrate the newly discovered patterns and customer segments with wide analyses of the socioeconomic situation of the area of interest, offering companies an useful tool to be consulted as assistance information for managers, according to the data-driven paradigm. The main advantage of the approach developed in this thesis is that it exploits free and publicly accessible datasets from national databases and open data sources to build a tool that can support retail managers in all marketing decisions, from locating the perfect spot for a new store to advertisement placement. To this extend, datasets on the Italian population are considered, and clustering algorithms are used to analyze the demographics of population, allowing to identify geo-referenced similarities. More in-depth analyses of the population distribution, positioning of stores, house prices and incomes distribution of Milan are also carried out, along with an analysis of the public transport of Lombardy. Analyzing the movements of people towards areas where stores are located allows to gain further insights: by integrating the public datasets of population and movements with real data on customers from stores it is possible to estimate the age group, reason for travel, mean of transport and origin of people entering a given store. It follows that a more precise identification of customers can result into specialized advertisement and/or a better located one. The results of these analyses have been made available in a website. All these insights can drive managers' decisions to a more data-driven and focused approach to marketing, as marketing efforts can be concentrated not only on specific customer segments generated by data mining tools but also on specific geographic areas.| File | Dimensione | Formato | |
|---|---|---|---|
|
2022_10_Bellini.pdf
non accessibile
Dimensione
3.42 MB
Formato
Adobe PDF
|
3.42 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/192116