Presentation on data preparation with pandas
- 1. MODY UNIVERSITY OF SCIENCE ANDTECHNOLOGY Colloquium Presentation CS 14.371 Submitted to- Dr. Pervesh Kumar Bishnoi Ms. Sonal Shukla Submitted by- Akshita Kanther B.Tech. IIIrd yr C2 Er.No.-180161
- 3. CONTENTS ļIntroduction ļWhy Should We Prepare Our Data ļPython ļPython Libraries ļPandas ļFeatures of Pandas ļCore Components Of Pandas ļPandas Operations ļTypical Pipeline For Data Preparation ļCommon Tasks Involved In Data Preparation ļApplications Of Pandas ļCompanies using Pandas ļSummary
- 4. INTRODUCTION Data preparation is the first step after you get your hands on any kind of dataset. This is the step when you pre-process raw data into a form that can be easily and accurately analyzed. Proper data preparation allows for efficient analysis - it can eliminate errors and inaccuracies that could have occurred during the data gathering process and can thus help in removing some bias resulting from poor data quality. Therefore a lot of an analyst's time is spent on this vital step.
- 5. WHY SHOULD WE PREPARE OUR DATA
- 6. ļ¼Garbage in, garbage out ļ¼Reduce errors ļ¼Remove duplicate records ļ¼Fix missing values ļ¼Correct range values ļ¼Fix formatting (i.e. date, text, number)
- 7. PYTHON ļ¶ Object-oriented, high-level programming language ļ¶ Used as a scripting language to connect existing components together ļ¶ Simple, easy to learn syntax emphasizes readability ļ¶ Supports modules and packages
- 8. PYTHON LIBRARIES Many popular Python toolboxes/libraries:- ā¢ NumPy ā¢ SciPy ā¢ Pandas ā¢ SciKit-Learn Visualization libraries:- ā¢ matplotlib ā¢ Seaborn
- 10. PANDAS ā¢ Pandas is a software library written for Python ā¢ Pandas has so many uses that it might make sense to list the things it can't do instead of what it can do ā¢ This tool is essentially your dataās home. Through pandas, you get acquainted with your data by cleaning, transforming, and analyzing it ā¢ Pandas is well suited for different kinds of data, such as: ļ¼ Tabular data with heterogeneously-typed columns ļ¼ Ordered and unordered time series data ļ¼ Arbitrary matrix data with row & column labels ļ¼ Unlabelled data ļ¼ Any other form of observational or statistical data sets To use the pandas library, you need to first import it. Just type this in your python console:
- 12. CORE COMPONENTS OF PANDAS The primary two components of Pandas are:- ļDataframe ļ Series A Series is essentially column and a Dataframe is a multidimensional Table made up of a collection of Series.
- 13. PANDAS OPERATIONS Using Python pandas, you can perform a lot of operations with series, data frames, missing data, group by etc. Some of the common operations for data manipulation are listed below:
- 14. TYPICAL PIPELINE FOR DATA PREPARATION
- 15. ā¢ The first step of a data preparation pipeline is to gather data from various sources and locations ā¢ Before any processing is done, we wish to discover what the data is about. At this stage, we understand the data within the context of business goals and Visualization of the data is also helpful here ā¢ The next stage is to cleanse the data of missing values and invalid values. We also reformat data to standard forms ā¢ Next we transform the data for a specific outcome or audience ā¢ We can enrich data by merging different datasets to enable richer insights ā¢ Finally, we store the data or directly send it out for analytics
- 16. COMMON TASKS INVOLVED IN DATA PREPARATION Tasks involved in Data Preparation Aggregation Augmentation Decomposing Deletion Blending Anonymization Data preparation involves one or more of the following tasks: ā¢Aggregation: Multiple columns are reduced to fewer columns. Records are summarized ā¢Anonymization: Sensitive values are removed for the sake of privacy ā¢Augmentation: Expand the dataset size without collecting more data. For example, image data is augmented via cropping or rotating ā¢Blending: Combine and link related data from various sources. For example, combine an employee's HR data with payroll data ā¢Decomposing: Decompose a data column that has sub-fields. For example, "6 ounces butter" is decomposed into three columns representing value, unit and ingredient ā¢Deletion: Duplicates and outliers are removed. Exploratory Data Analysis (EDA) may be used to identify outliers
- 19. SUMMARY Raw data is usually not suitable for direct analysis. This is because the data might come from different sources in different formats. Moreover, real-world data is not clean. Some data points might be missing. Some others might be out of range. There could be duplicates. Data preparation is therefore an essential task that transforms or prepares data into a form that's suitable for analysis. Data preparation assumes that data has already been collected. However, others may consider data collection and data ingestion as part of data preparation. Within data preparation, it's common to identify sub-stages that might include data pre- processing, data wrangling, and data transformation.