| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.25 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Na Economia atual, obter informação com qualidade e atempadamente é uma mais valia nas organizações para que estas alcancem uma vantagem competitiva, principalmente quando se trata de grandes empresas com portefólios de negócios alargados. O propósito desta dissertação consiste na implementação de processos de BI em ferramentas de Big Data, de forma a comprovar a sua mais-valia no auxílio à tomada de decisão. No âmbito da implementação do projeto Profit & Loss Comercial, na Sonae MC, mais especificamente BIT, foi identificada a necessidade de alterar o atual processo de disponibilização de receitas comercias, mais especificamente as relativas com processos marginais e promocionais. O pedido advém da necessidade de alterar o comportamento do processo atual, de forma a permitir uma visão analítica da informação financeira atualmente existente. É então intuito distribuir os valores de receitas cadastrados pelos utilizadores, de forma a que estes espelhem o seu impacto na rentabilidade comercial diária de toda a organização. Para tal, foram definidos algoritmos de distribuição juntamente com a equipa de negócio Sonae, que visam ratear as receitas cadastradas de acordo com a disponibilidade do comercial, de uma forma equilibrada pelos dias de afetação da receita e peso diário de cada loja na empresa como um todo. Os resultados obtidos possuem um grau de confiança e sucesso grande pois os processos implementados cumprem todos os requisitos e com um desempenho excelente. Logicamente que para atingir tempos aceitáveis foi necessário afinar os processos através de algumas técnicas, havendo ainda assim a necessidade de realizar uma verificação e melhoria constante.
In the current economy, having information with quality and timely is a profit in organizations so they can achieve a competitive advantage, especially when it comes to large companies with extended portfolios. The purpose of this thesis is to implement BI processes in Big Data tools to prove its added value in aiding decision making. As part of Commercial Profit & Loss implementation in Sonae MC, more specifically BIT, we identified the need to change the current process of making commercial revenues, namely those related to marginal and promotional processes. The first step of this implementation was the characterization and specification of the problem, thus realizing the real needs and requirements of the commercial teams. Subsequently, the processes were implemented, executed and validated in the existing Big Data tools on the organization. At last, some conclusions were announced such as questions for further investigation. The request comes from the need to change the behavior of the current process, to add an analytical view of the financial information that currently exists. It is then intended to distribute the revenue values registered by users, so they reflect their impact on the daily business profitability of the entire organization. To this end, distribution algorithms were defined together with the Sonae business team, which aim to split the revenues registered according to the availability of the commercial, in a manner balanced by the days of revenue allocation and daily weight of each store in the company. The results obtained have a degree of confidence and great success because the processes implemented meet all the requirements and with an excellent performance. Of course, to achieve acceptable times it was necessary to fine-tune the processes through some techniques, but there is still a need for constant verification and improvement.
In the current economy, having information with quality and timely is a profit in organizations so they can achieve a competitive advantage, especially when it comes to large companies with extended portfolios. The purpose of this thesis is to implement BI processes in Big Data tools to prove its added value in aiding decision making. As part of Commercial Profit & Loss implementation in Sonae MC, more specifically BIT, we identified the need to change the current process of making commercial revenues, namely those related to marginal and promotional processes. The first step of this implementation was the characterization and specification of the problem, thus realizing the real needs and requirements of the commercial teams. Subsequently, the processes were implemented, executed and validated in the existing Big Data tools on the organization. At last, some conclusions were announced such as questions for further investigation. The request comes from the need to change the behavior of the current process, to add an analytical view of the financial information that currently exists. It is then intended to distribute the revenue values registered by users, so they reflect their impact on the daily business profitability of the entire organization. To this end, distribution algorithms were defined together with the Sonae business team, which aim to split the revenues registered according to the availability of the commercial, in a manner balanced by the days of revenue allocation and daily weight of each store in the company. The results obtained have a degree of confidence and great success because the processes implemented meet all the requirements and with an excellent performance. Of course, to achieve acceptable times it was necessary to fine-tune the processes through some techniques, but there is still a need for constant verification and improvement.
Description
Keywords
Business Intelligence Data Warehouse Big Data Hadoop MapReduce Spark Hive Scala