July 2021 Summaries
6 posts from Census
Filter
Month:
Year:
Post Summaries
Back to Blog
The Census team has rebranded with a new identity that reflects their enthusiasm for empowering teams and people with new tools and frameworks. The new logo features a stylized hexagon symbol with stripes of movement, representing data in motion and efficiency. The brand's colors, typography, and overall aesthetic are inspired by the Vaporwave and Bauhaus art movements, conveying a sense of vibrancy, simplicity, and trustworthiness. The new identity aims to unite behind a fresh start for the Census brand and the data community, while also honoring its history and roots.
Jul 28, 2021
1,058 words in the original blog post.
Lead scoring is an important tool for sales and marketing teams, allowing them to focus on the most engaged users and convert them at a higher rate. It works by using rules to "qualify" each lead in the pipeline, assigning it a new priority based on its historical conversion rates. To get started with lead scoring, a basic system requires a data stack, including collection, storage, modeling, and activation tools. The quality of the data is crucial, as garbage-in-garbage-out applies, and ensuring that data is stored in a standardized way, regularly snapshotted, and uniquely matched to conversions can make a significant difference. Lead scoring models can be used for marketing email automation and product qualified leads (PQLs), which convert at a higher rate and are less likely to churn once they become customers. Effective lead scoring systems require an operational analytics stack, including data collection, storage, modeling, and sharing tools. By prioritizing the right leads and automating engagement, sales teams can save valuable time and improve conversion rates.
Jul 21, 2021
1,602 words in the original blog post.
Snowflake users are looking for ways to export their data from Snowflake in CSV format to utilize it with other tools and platforms. The potential for data expansion is shifting faster than typing a WHERE clause, making data management more complex. Users have three main motivations for exporting their CSV files: to import into other tools, move data across cloud platforms, or visualize data for analysis and reporting due to network policy limitations. Snowflake provides convenient features for data analysis, but users can also use various methods to export CSV data, including using a COPY command with cloud storage, SnowSQL, or BI tools like Mode or Metabase. However, Census is considered the "right" way to move data out of Snowflake due to its ability to sync data between Snowflake and other popular destinations in real-time.
Jul 15, 2021
1,329 words in the original blog post.
This summary provides an overview of the article's content on marketing mix modeling using Python. The key points covered include: Marketing mix models are used to determine market attribution, measuring the impact of each marketing channel, and are useful for understanding relationships between channels and target metrics, distinguishing high ROI channels from low ones, and predicting future conversions. A marketing mix model is built in 4 steps: importing libraries and data, performing exploratory data analysis (correlation matrices, pair plots, feature importance), building the model using ordinary least squares regression, and plotting actual vs predicted values. The output of a marketing mix model provides insights such as the proportion of variation explained by the model, p-values for each predictor, and can help improve sales by identifying significant predictors of the target variable.
Jul 12, 2021
1,443 words in the original blog post.
You can leverage HubSpot's flexibility to integrate data from other systems into it, even if the data doesn't fit HubSpot's default properties. To do this, you can create a custom contact property using the HubSpot API and Python. First, you need to get set up with the HubSpot API and Python by connecting to the API with your API key and installing the necessary libraries. Next, you check if the property already exists in the HubSpot web interface or through the API. If it doesn't exist, you create a new custom contact property using the API or web interface. Once the property is created, you can update its value by identifying the contact's id and writing data to the custom property. Finally, some general advice on the HubSpot API includes implementing error handling for rate limits and daily API usage.
Jul 07, 2021
1,590 words in the original blog post.
Reverse ETL (Extract, Transform, Load) involves moving data from a data warehouse into tools like Google Sheets for analysis and collaboration. This process can be simplified by using libraries such as psycopg2, pandas, gspread, and gspread-dataframe to connect to Postgres databases and write data to Google Sheets. To set up access to Google Sheets, users need to create a GCP account, enable the Google Drive API, and generate credentials for server-to-server access. Once these pieces are in place, users can use Python scripts to read from Postgres, retrieve data, and write it to Google Sheets. However, as the data flow becomes more complex, issues such as custom code and scaling problems may arise, highlighting the need for integrations that can scale with the data volume and user needs.
Jul 01, 2021
1,071 words in the original blog post.