Find Ideal Location for Business in Bangladesh

The dataset has 21 columns that carry the features (questions) of 988 respondents. The efficiency of any machine learning model is heavily dependent on its raw initial dataset. For this, we had to be extra careful in gathering our information. We figured out that for our particular problem, we had to go forward with data that was not only authentic but also versatile enough to get the proper information from relevant sources. Hence we opted to build our dataset by dispatching a survey questionnaire among targeted audiences. Firstly, we built the questionnaire with inquiries that were made after keen observation. Studying the behavior from our intended audience, we came up with factual and informative queries that generated appropriate data. Our prime audience were those who were highly into buying fashion accessories and hence we had created a set of questionnaires that emphasized on questions related to that field. We had a total of twenty one well revised questions that gave us an overview of all answers that were going to be needed within the proximity of our system. As such, we had the opportunity to gather over half a thousand authentic leads and concluded upon our initial raw dataset accordingly.本数据集共计21个特征列，每列代表一个调研问题，涵盖988名受访者针对该问题的应答数据。机器学习模型的建模性能在极大程度上取决于其原始初始数据集的质量。为此，我们在数据采集过程中始终保持高度严谨。经分析，针对本次研究的特定任务，我们需选用兼具真实性与多样性的数据，以从相关数据源中获取精准有效的目标信息。因此，我们决定通过向目标受众发放调研问卷的方式构建本数据集。首先，我们通过细致的观察与调研设计了问卷的提问内容。通过研究目标受众的行为特征，我们设计出兼具客观性与信息价值的调研问题，以确保采集到适配需求的数据。本次调研的核心受众为热衷购买时尚配饰的群体，因此我们设计的问卷聚焦该领域的相关问题。我们最终确定了21道经过多轮修订完善的调研问题，这些问题覆盖了本系统研发所需的全部应答维度。基于此，我们最终成功收集到500余份真实可信的受访者数据，并据此完成了初始原始数据集的构建工作。