shak9345/airbnb_New_York_prices

<video src="https://huggingface.co/datasets/shak9345/airbnb_New_York_prices/resolve/main/Presentation.mp4" controls="controls" style="max-width: 720px;"></video> # Airbnb NYC Price Analysis ## Project Overview This project explores Airbnb listings in New York City in order to understand which factors are most closely related to listing prices. The analysis is based on exploratory data analysis (EDA) and focuses on identifying patterns between listing price and variables such as location, room type, reviews, minimum nights, and host-related characteristics. ## Research Question **What factors influence Airbnb prices in New York City?** ## Dataset The dataset contains around **49,000 Airbnb listings** and **16 variables**. It includes both numerical and categorical data, such as: - `price` - `room_type` - `neighbourhood_group` - `neighbourhood` - `minimum_nights` - `number_of_reviews` - `reviews_per_month` - `availability_365` - `latitude` - `longitude` This combination makes the dataset suitable for exploring how different listing characteristics are associated with Airbnb prices. ## Data Quality and Cleaning Before beginning the analysis, the dataset was inspected for missing and invalid values. Main cleaning decisions: - Missing values in `reviews_per_month` were filled with **0** - Listings with `price = 0` were treated as unrealistic and replaced with the **mean valid price** - Identifier-like columns such as `id`, `host_id`, `name`, and `host_name` were removed from the analytical dataset - Because the `price` variable contained many extreme outliers, most of the visual analysis focused on listings up to the **95th percentile** of price At first, the 99th percentile was considered, but the plots were still difficult to interpret due to the large number of extreme outliers. The 95th percentile provided clearer and more readable visualizations. ## Main Insights The analysis suggests that **location** and **room type** are the strongest factors associated with Airbnb prices in New York City. Key findings: - Listings in **Manhattan** tend to be more expensive than listings in the other boroughs - Certain neighborhoods such as **Tribeca, NoHo, and DUMBO** show especially high average prices - **Entire homes/apartments** are significantly more expensive than **private rooms** and **shared rooms** - The combination of **location and room type** explains price differences better than either variable alone - **Minimum nights** show some relationship with price, but the effect is weaker - **Number of reviews** does not appear to strongly increase price - Most numerical variables have only **weak linear correlations** with price ## Key Visualizations The notebook includes several visualizations used to explore the research question, including: - Room type distribution - Neighborhood group distribution - Price distribution - Price by room type - Price distribution by neighborhood group - Average price by neighborhood and room type - Geographical distribution of prices - Top 10 most expensive neighborhoods - Correlation matrix ## Example Visualizations ### Price by Room Type  This visualization shows a clear difference between room types. Entire homes/apartments have the highest prices, private rooms are cheaper, and shared rooms are usually the cheapest. ### Price Distribution by Neighborhood Group  This graph shows that location matters strongly. Manhattan has the highest overall price distribution, while the Bronx and Staten Island tend to have lower price levels. ### Average Price by Neighborhood and Room Type  This plot highlights that price is influenced by the combination of neighborhood and room type, not by only one factor. ## Conclusion Based on the analysis, the main factors influencing Airbnb prices in New York City are **location** and **room type**. Listings in more central and high-demand areas tend to be more expensive, and listings offering an entire home or apartment are priced significantly higher than private or shared-room options. Other variables, such as minimum nights, number of reviews, and availability, appear to have weaker or less consistent relationships with price. Overall, the dataset suggests that Airbnb prices in New York are driven mainly by spatial and structural listing characteristics. ## Limitations This is an exploratory analysis, so the results show **associations rather than causal relationships**. Some limitations of the dataset: - It does not include apartment size, number of bedrooms, amenities, or interior quality - It does not capture seasonal demand or local tourism events - Some conclusions may be affected by variables that are not available in the dataset ## Future Work Future work could extend this analysis by: - building predictive models for Airbnb price - testing feature importance more formally - enriching the dataset with additional property-level features such as number of bedrooms, amenities, or distance from major attractions<video src="https://huggingface.co/datasets/shak9345/airbnb_New_York_prices/resolve/main/Presentation.mp4" controls="controls" style="max-width: 720px;"></video> # 爱彼迎（Airbnb）纽约房价分析 ## 项目概述 本项目针对纽约市的爱彼迎（Airbnb）房源展开探索，旨在明确与房源价格高度相关的核心影响因素。本次分析基于探索性数据分析（Exploratory Data Analysis，EDA），重点挖掘房源价格与区位、房源类型、评论信息、最短入住时长以及房东特征等变量间的关联模式。 ## 研究问题 **纽约市爱彼迎房源价格受哪些因素影响？** ## 数据集 本数据集包含约49000条爱彼迎房源数据与16项变量，涵盖数值型与分类型两类数据，具体包括： - `price`（价格） - `room_type`（房源类型） - `neighbourhood_group`（行政区群组） - `neighbourhood`（街区） - `minimum_nights`（最短入住时长） - `number_of_reviews`（评论数量） - `reviews_per_month`（月均评论数） - `availability_365`（年度可预订天数） - `latitude`（纬度） - `longitude`（经度） 该数据集的变量组合使其适用于探索不同房源特征与爱彼迎房价间的关联关系。 ## 数据质量与清洗 正式开展分析前，本团队已对数据集进行缺失值与无效值排查。 主要清洗决策如下： - 将`reviews_per_month`的缺失值填充为**0** - 将价格为0的房源视为不合理数据，替换为有效价格的**均值** - 移除分析数据集中诸如`id`、`host_id`、`name`与`host_name`这类标识符类列 - 由于`price`变量存在大量极端异常值，多数可视化分析仅聚焦于价格处于95%分位数以内的房源。最初曾考虑使用99%分位数，但因极端异常值过多导致图表难以解读，最终选用95%分位数以获得更清晰易读的可视化效果。 ## 核心发现 分析结果表明，**区位**与**房源类型**是影响纽约市爱彼迎房源价格的最核心因素。 关键发现如下： - 曼哈顿（Manhattan）的房源价格普遍高于其他行政区 - 翠贝卡（Tribeca）、诺霍（NoHo）与丹波（DUMBO）等特定街区的平均房价尤为高昂 - **整套公寓/住宅**的价格显著高于**私人房间**与**共享房间** - 区位与房源类型的组合相比单一变量，能更好地解释房价差异 - 最短入住时长与价格存在一定关联，但影响较弱 - 评论数量似乎并未显著推高房价 - 多数数值型变量与价格仅存在微弱的线性相关性 ## 关键可视化内容 本分析文档包含多个用于探索研究问题的可视化图表，具体包括： - 房源类型分布 - 行政区群组分布 - 价格分布 - 按房源类型划分的价格 - 按行政区群组划分的价格分布 - 按街区与房源类型划分的平均价格 - 价格的地理分布 - 十大最贵街区 - 相关系数矩阵 ## 示例可视化 ### 按房源类型划分的价格  该可视化清晰展现了不同房源类型间的价格差异：整套公寓/住宅的价格最高，私人房间次之，共享房间通常最为低廉。 ### 按行政区群组划分的价格分布  该图表直观体现了区位的重要影响：曼哈顿的整体房价分布最高，而布朗克斯与斯塔滕岛的房价水平普遍较低。 ### 按街区与房源类型划分的平均价格  该图表凸显了房价同时受街区与房源类型双重因素影响，而非单一变量决定。 ## 研究结论 基于本次分析，影响纽约市爱彼迎房源价格的核心因素为**区位**与**房源类型**。位于核心区域与高需求地段的房源价格普遍更高，提供整套住宅或公寓的房源定价显著高于私人房间或共享房间选项。其余变量，如最短入住时长、评论数量与可预订天数，与房价的关联度较弱且一致性不足。总体而言，本数据集表明纽约市爱彼迎房源价格主要由空间属性与房源结构特征决定。 ## 研究局限 本研究属于探索性分析，因此所得结果仅展示**关联关系而非因果关系**。 本数据集存在以下局限： - 未涵盖公寓面积、卧室数量、配套设施或内饰品质等信息 - 未纳入季节性需求或本地旅游活动相关数据 - 部分结论可能受到数据集中未包含的变量影响 ## 后续研究方向 后续研究可从以下方向拓展本次分析： - 构建爱彼迎房源价格预测模型 - 更正式地开展特征重要性测试 - 补充更多房源层面的特征以丰富数据集，例如卧室数量、配套设施或与主要景点的距离