Demonstration · Illustrative data · BI Consulting Services
Comfort Keepers · Hazleton, PA  × BI Consulting Services

Home-care operations & HR board
every hour scheduled, every hour delivered

One board for the Hazleton office and the eight communities it serves across Luzerne, Schuylkill and Columbia counties — caregiver hiring and turnover, scheduled vs delivered hours, client census by county, missed and late visits, overtime and payer mix. Illustrative numbers, real structure.
Prepared for Louis
Director, Human Resources
September 2026
County
Period
Showing all 8 service areas · · deltas vs the prior equal-length period

① TrendScheduled vs delivered hours, by week

Bars = hours scheduled; filled bars = hours actually delivered. Gap = unfilled shifts.
ScheduledDeliveredFill rate (%)

② AnswersWhat this board answers

Three questions from the discovery call, each answered from the filtered data.

③ SignatureClient census across the three-county service area

Bubble = active clients this week (latest week in range); ring color = fill rate for that community. Office at 261 S Church St, Hazleton.
Clients (area ∝ census)Fill ≥ 95%92–95%< 92%Hazleton office

④ HRCaregiver hiring funnel & turnover

Applicants through to first shift worked, summed over the period.
Annualized turnover
Median days, applicant → first shift
Caregivers who started (up) vs caregivers who left (down), by week

⑤ SchedulingWeekly coverage grid — fill rate by day & shift

Each cell is delivered ÷ scheduled hours for that day and shift, averaged across the period. Darker = better covered.

⑥ RankingMissed & late visits per 100 scheduled, by community

Missed = no caregiver arrived; late = more than 15 minutes after the scheduled start.

⑦ Payer mixClient census by payer source

Share of active clients this week, by who pays for care.

⑧ Ask the dataAI layer on top of the board

A natural-language question, answered from the same governed dataset with a chart reference. Mock-up of the conversational layer.
Also from BI Consulting Services

Private AI on your own server

The "ask the data" ideas on this board are designed to run on a GPU server inside your own network — an open-weight model over your own documents, cited answers, no per-seat licence, nothing leaving the building. The short deck below explains how it works. Scroll through, or download it as a PDF.

Private AI Servers — slide 1 of 14
Private AI Servers — slide 2 of 14
Private AI Servers — slide 3 of 14
Private AI Servers — slide 4 of 14
Private AI Servers — slide 5 of 14
Private AI Servers — slide 6 of 14
Private AI Servers — slide 7 of 14
Private AI Servers — slide 8 of 14
Private AI Servers — slide 9 of 14
Private AI Servers — slide 10 of 14
Private AI Servers — slide 11 of 14
Private AI Servers — slide 12 of 14
Private AI Servers — slide 13 of 14
Private AI Servers — slide 14 of 14

Questions? Book 30 minutes: calendly.com/powerbiconsultingservices/30min