Challenge: The client engaged SSI to optimize and automate its end-to-end business operations, which included product marketing, online sales, payment processing, order fulfillment, contract management, customer relationships, user experience, product development, and dataset onboarding.
Workflow Fragmentation
The business workflow was fragmented, requiring manual processing by the back office. This caused order processing lag for global sales originating from different time zones.
Delayed Dataset Onboarding Affecting Sales
The speed of dataset onboarding did not meet customer demand. With new datasets being provided by data providers as they became available, the client was experiencing lost sales
Solution: SSI developed significant enhancements to increase the speed and scale of the platform; read the details below:
E-commerce Platform
SSI has developed a self-service e-commerce platform for the client, enabling its customers to license and buy datasets and technology services. This has been implemented as a customized version of Shopify to combine out-of-the-box features with unique digital product features. The platform provides an end-to-end e-commerce workflow that includes:
- Product catalog management and synchronization
- Shopping cart with tiered/targeted pricing by customer
- Legal agreements covering multi-level contracts (master, vendor, dataset)
- Account management with billing/charges for manual and recurring invoices
- Integration with Oracle NetSuite for financial reporting and reconciliation
- Search analytics and traffic tracking
- Integration with LDAP and HubSpot
- Integration with Node.js backend to keep track of the Shopify store
- A fully customized data signals site in WordPress for routing sales to the Shopify store
Dataset Onboardings
SSI’s data engineering team performed rapid onboarding of vendor datasets to the backend datastore. According to the structure and quality of incoming data, relational (SQL) and programmatic (Python) techniques were used. While peers in the industry spend months, sometimes longer, onboarding datasets, the SSI team does it in hours using proprietary data fabric and advanced ETL methodologies. Additionally, SSI has built an automated tool for faster onboarding of simple datasets.
Platform Development
- Migrated server-side architecture to Node.js for scalability of I/O intensive data requests.
- Enhanced caching capabilities to boost the client-side performance of visualization components showing large volumes of data.
- Integrated HighCharts library for JS-based charting to deliver a responsive click-through UX/UI for visualization components.
- Integrated the platform with APIs of third-party CRM HubSpot.
- Created widgets runnable from different environments (like C# and Python) for tight integration with legacy client apps.
Machine Learning Prospects
SSI performed a constructive POC & Demo regarding Sentiment Analysis on Moody’s Stock News Data, revealing a significant correlation between stock prices, volumes, and sentiment analysis results.
Abstract
The comprehensive analysis comprised of stock news sentiment using two sentiment analysis tools: VADER (Valence Aware Dictionary for Sentiment Reasoner) and Text Blob. The study was conducted in three phases:
Data acquisition
We pulled data from Moody’s News Live servers for specific date ranges.
Preprocessing
Python’s NLTK was used to clean and preprocess the data into meaningful insights, such as identifying named entities (e.g., names of people, organizations, and locations).
Sentiment Analysis
SSI used two popular tools for the analysis, using a dataset of 1000 rows from popular tickers like Uber, Dell, and Facebook. The data was cleaned using Python’s NLTK preprocessing. The correlation between stock prices, volumes, and sentiment analysis results is particularly evident when examining specific time frames and symbols, with negative sentiment aligning with price drops and positive sentiment correlating with price increases.
The study concluded that Moody’s sentiment analysis correlates with stock prices and volumes, further underscoring the potential predictive power of sentiment analysis in the finance and investment industries.
Result: Through cross-platform integrations, SSI has eliminated back-office processing from customers’ online signup through order processing to datasets delivery. This has provided a seamless onboarding experience for the global clientele. The SSI team has already onboarded 250 datasets for the client in less than a year. This is an average of 20+ datasets per month, which is about ten times more than the throughput of competitors.
Tools & Technologies:
