database

1. Data We Collect

PharmDex collects only the minimal data necessary to function, maintain security, and optimize performance. We categorize the collected information into three clear areas:

User Search Views & Input

When you execute search queries, we collect search query parameters, molecular structures (such as API and Excipient SMILES representations), and related metadata. This allows our backend to perform computational predictions and cache results to accelerate future lookups.

Location Statistics

To understand global adoption and allocate computing resources optimally, we parse incoming HTTP request IP addresses to resolve approximate location data (country and town only). We do not store or track exact IP coordinates.

insights

2. How We Use Information

We strictly limit the use of collected details to enhancing your direct workspace experience and scientific accuracy:

check_circlePrediction Caching: Pre-calculating and saving complex 3-Tier ML models outputs for frequently viewed API-excipient SMILES pairs.
check_circleAggregated Scientific Statistics: Visualizing regional drug discovery research trends on an anonymized scale using town/country metrics.
check_circleRate Limiting & Threat Prevention: Shielding our API endpoints from abusive traffic and excessive load.

psychology

3. Feedback Telemetry & AI Training

When you submit qualitative remarks or rating points via the Beta Feedback Sidebar, we capture explicit telemetry:

This telemetry includes your score (1 to 5), qualitative observation text (e.g. noted physical incompatibilities or literature disagreements), prediction metrics, and active ML routing tier identifiers.

Model Optimization Notice

We use this valuable telemetry to continuously train, fine-tune, and optimize our Ensemble-Based and Rule-Based chemical compatibility prediction models. Your real-world experimental feedback is crucial for ensuring the platform maintains and elevates its scientific accuracy over time.

shield

4. Data Security & Storage

We employ strict security practices to keep your session logs and queries secure:

All communications are encrypted transit-wide via HTTPS/TLS protocols.
Data databases are deployed within standard VPC environments in GCP.
Access controls are restricted strictly to authorized systems and principal investigators of the Dr. PP Research Lab.
We never share, lease, or sell any user information or searched molecules to external third-party entities.

manage_accounts

5. Your Choices & Deactivation

We provide you with clear control over your profile and stored parameters:

Self-Service Account Deactivation

If you choose to suspend your account, you can do so directly from your User Profile Page. Consistent with standard academic integrity procedures, we handle deactivations via soft deactivation. Your profile state will be deactivated immediately, and you will be securely logged out. However, historical, anonymized search logs and contributed molecular compatibility values are retained to maintain database consistency for the scientific fraternity.