# Dashboards and analytics Kit: Apache projects for small business
October 29, 2026. [Read the companion article](https://jwatte.com/blog/apache-dashboards-analytics-small-business/).
## How to use this kit
Write the question your report must answer and name the person who owns its definition. Every server project here retains the IT tier; the chart libraries belong with your developer. Start with a read only sample, then test access and recovery. No installation or performance testing was done for this research.
## Provider cards
### Superset
Use it for shared database dashboards, such as a two location bakery comparing waste or a nine person HVAC contractor checking uninvoiced jobs. The [FAQ](https://superset.apache.org/docs/faq/) reports community sizing of 8 GB RAM and two vCPUs, not a universal minimum. Our pilot assumes one 4 vCPU/8 GiB host. The [Compose guide](https://superset.apache.org/docs/installation/docker-compose/) rejects its examples as production ready; follow the [current Kubernetes guidance](https://superset.apache.org/docs/installation/kubernetes/) matched to the installed version. The [distribution](https://downloads.apache.org/superset/) lists 6.1.0 with a May 13, 2026 file date.
Ask who owns database permissions, the metadata backup and restore, chart definitions and update testing. [Preset Professional](https://www.preset.io/pricing/) is $25/user/month monthly or $20 paid annually; Starter is free for up to five users. Prices read October 29, 2026. First week: agree one metric, publish one chart, test a restricted viewer, export definitions and restore the metadata database. Use 24 to 60 setup hours and 6 to 16 admin hours/month as our IT oversight assumption, excluding data engineering.
### Druid
Use it for rapid event queries, such as a thirty person delivery company investigating missing scans or a forty person ecommerce team studying nonpayment checkout errors. Check PCI DSS before collecting payment data. The [local tutorial](https://druid.apache.org/docs/latest/tutorials/) requires at least 6 GiB RAM, Unix like OS, Java 17, Python 3 and Perl 5. Our 4 vCPU/8 GiB host is an evaluation assumption. The [single server guide](https://druid.apache.org/docs/latest/operations/single-server/) deprecates older startup profiles. [Release 38.0.0](https://druid.apache.org/downloads/) is dated October 1, 2026.
Ask how duplicate and late events are handled, how counts reconcile and how data is rebuilt. [Imply Polaris](https://www.imply.io/pricing/) Starter begins at $100/month for a limited proof of concept; production Standard starts at $600/month, with usage costs additional to a simple starting comparison. First week: load a bounded synthetic sample, compare counts, simulate missing and duplicate events, restore and requery. Our 24 to 60 setup hours and 6 to 16 monthly hours cover evaluation operations, not a resilient production platform.
### Zeppelin
Use it when a skilled analyst needs a notebook, such as a seven person print shop reviewing material use or a ten person farm and winery comparing batch loss. The [install guide](https://zeppelin.apache.org/docs/latest/quickstart/install.html) lists JDK 11 and tested operating systems, with no universal hardware floor. Our small SQL/file pilot assumes one 4 vCPU/8 GiB host and no external compute cluster. Use the current [0.12.1 download](https://zeppelin.apache.org/download.html), dated June 12, 2026, rather than old image tags in examples.
Ask who can execute code, which interpreters are enabled and whether notebooks, configuration and dependencies restore together. [Power BI Pro](https://www.microsoft.com/en-us/power-platform/products/power-bi/pricing) is a narrower report sharing alternative at $14/user/month paid yearly. First week: save a sample input, reproduce a calculation after restart and have a second person explain it. Our operating assumption is 4 to 12 setup hours and 1 to 3 monthly hours, excluding analytical work.
### Cloudberry (Incubating)
Needs distributed database skills. [Production recommendations](https://cloudberry.apache.org/docs/deployment/platform-requirements/) include two 16+ core/32+ GB coordinators and two 8+ core/32+ GB segments; a one node test is not the production minimum. A forty five person wholesaler with existing data staff might justify it. Ask for node sizing, failure tests and an export plan. Assume 6 to 16 monthly operational hours, excluding engineering. BigQuery is a managed warehouse alternative; see its usage price below.
### Doris
Needs frontend/backend operations and ingestion ownership. Its [4.x requirements](https://doris.apache.org/docs/4.x/install/preparation/env-checking/) distinguish co-located tests from production recommendations of separated services and 16+ core/64+ GB machines. A thirty person delivery company is an illustration. Ask for actual node count, recovery and ingestion costs; assume 6 to 16 monthly operational hours. [VeloDB Standard US East](https://www.velodb.io/pricing) lists $0.134/vCPU/hour compute, $0.00005/GB/hour storage and $0.000332/GB/hour cache. Do not present those units as an all inclusive subscription.
### Kylin
Needs existing data platform staff. The [prerequisites](https://kylin.apache.org/docs/deployment/prerequisite/) recommend 16 vCore, 64 GB RAM and 500 GB for the server plus Hadoop services. Even [single node mode](https://kylin.apache.org/docs/deployment/single_node_mode/) depends on additional infrastructure. A twenty five person manufacturer with a warehouse might justify it. Ask which dependencies are already paid for and who maintains them. Assume 6 to 16 monthly operational hours, plus engineering. BigQuery is an alternative approach, not hosted Kylin.
### Pinot
Needs streaming and distributed system skills. [Deployment docs](https://docs.pinot.apache.org/operate-pinot/deployment) describe controller, broker, server, a recommended three node ZooKeeper ensemble for production and durable deep storage. No universal hardware floor is published. A forty person ecommerce business building customer dashboards is an illustration. Ask about authentication, retention, partitioning and failure recovery. Assume 6 to 16 operational hours/month plus integration. StarTree lists $0.21 per production vCPU hour fully managed, or $0.11 per production vCPU hour in your own cloud account plus infrastructure (list prices, October 29, 2026); get a workload quote. BigQuery may cover a narrower reporting need.
[BigQuery Iowa pricing](https://cloud.google.com/bigquery/pricing), read October 29, 2026, is $6.25/TiB of on demand query processing after the first 1 TiB/month free. Storage is additional. Fill in scanned volume and storage before comparing costs. Incubating means [not yet fully endorsed by the ASF](https://incubator.apache.org/policy/incubation.html).
## Full project table
| Project | What it does for a business | Who runs it | Latest recorded release and date |
|---|---|---|---|
| [Apache Cloudberry (Incubating)](https://cloudberry.incubator.apache.org/) | Spreads analytical database queries across servers for larger reporting workloads. | Your IT department or provider | 2.1.0-incubating; 2026-04-15 |
| [Apache DataSketches](https://datasketches.apache.org) | Estimates counts and patterns in large datasets without keeping every detail. | Your developer | 0.6.0; 2026-10-26; Rust Core component |
| [Apache Doris](https://doris.apache.org/) | Stores business data for fast reports and interactive analysis. | Your IT department or provider | 4.1.5; 2026-10-15 |
| [Apache Druid](https://druid.apache.org/) | Answers fast analytical queries about events, activity and trends. | Your IT department or provider | 38.0.0; 2026-10-01 |
| [Apache ECharts](https://echarts.apache.org) | Adds interactive charts to websites and business applications. | Your developer | 5.6.0; 2024-12-28; registry |
| [Apache Flagon](https://flagon.apache.org) | Helps developers study how people use software interfaces. | Your developer | 2.4.0; 2024-04-03; file date; UserALE.js component |
| [Apache Kylin](https://kylin.apache.org) | Prepares large datasets so business reports can answer queries quickly. | Your IT department or provider | 5.0.5; 2026-07-13 |
| [Apache Ossie (Incubating)](https://ossie.incubator.apache.org/) | Defines a shared format for business metrics so different reporting tools can interpret them consistently. | Your developer | Not listed; date not listed |
| [Apache Pinot](https://pinot.apache.org) | Feeds dashboards and applications with fast analysis of incoming events. | Your IT department or provider | 1.6.0; 2026-10-28; registry |
| [Apache Superset](https://superset.apache.org) | Builds shared dashboards and charts from business databases. | Your IT department or provider | 6.1.0; 2026-05-13; file date |
| [Apache Zeppelin](https://zeppelin.apache.org/) | Provides shared notebooks for exploring data with queries, code and charts. | Your IT department or provider | 0.12.1; 2026-06-12; registry |
Registry labels refer to the [ASF project directory data](https://projects.apache.org/json/foundation/projects.json), read October 29, 2026 at 21:00 UTC. Component, preview and file date qualifications are part of the release description.
## Cost worksheet
Low/high users: ___ / ___. Low/high nodes: ___ / ___. Monthly server total: ___ / ___. Retained backup GB: ___ / ___. Setup hours: ___ / ___. Admin hours/month: ___ / ___. Managed subscription and billing term: ___. Usage charges and implementation quote: ___.
Setup = hours × $68.09. Monthly = hosting + backups + admin hours × $68.09. Three year total = setup + 36 × monthly. Savings = alternative total minus self hosted total. If alternative monthly exceeds self hosted monthly, break even = round up((self setup minus alternative setup)/(alternative monthly minus self monthly)); otherwise never unless an initial setup advantage changes the result.
Shared inputs read October 29, 2026: [DigitalOcean](https://www.digitalocean.com/pricing/droplets) $48/month for 4 vCPU/8 GiB; [B2](https://www.backblaze.com/cloud-storage/pricing) $6.95/TB/month with 10 GB free; [BLS wage](https://data.bls.gov/oesprofile/?areas=INDUSTRY,STATE,MSA&major_group=150000&measure=01&occupation=151244&year=2025) $47.66 divided by [compensation share](https://data.bls.gov/timeseries/CMU2020000000000P) 0.700, rounded to $68.09/hour. A provider will charge more. All prices are USD before tax. Retained backup volume is 100 to 500 GB, with zero chargeable transfer/API usage assumed. Each example uses one host in both cases. Source database cost, custom development, migration, training, compliance and incidents are excluded equally and need quotes.
| Superset pilot versus Preset Professional | Low case | High case |
|---|---:|---:|
| Our assumption: setup hours | 24 | 60 |
| Our assumption: administration hours/month | 6 | 16 |
| Our assumption: stored backup gb | 100 | 500 |
| Servers/month | $48 | $48 |
| Setup labor | $1,634 | $4,085 |
| Total/month | $457 | $1,141 |
| Self hosted, 36 months | $18,092 | $45,156 |
| Preset monthly plan, monthly | $125 | $250 |
| Alternative, 36 months | $4,500 | $9,000 |
| Savings from self hosting, 36 months | Negative $13,592 | Negative $36,156 |
| Break even month | Never | Never |
Setup = hours × $68.09. Monthly = servers + (stored GB minus 10) × $0.00695 + administration hours × $68.09. Total = setup + 36 × monthly; savings = alternative total minus self hosted total. Calculations use unrounded values; displayed totals round to whole dollars.
Our assumption: five to ten paid users, no embedded viewer add-on, zero separately charged alternative setup. Managed Professional is cheaper; check whether free Starter already meets the need.
| Druid private evaluation operating budget | Low case | High case |
|---|---:|---:|
| Our assumption: setup hours | 24 | 60 |
| Our assumption: administration hours/month | 6 | 16 |
| Our assumption: stored backup gb | 100 | 500 |
| Servers/month | $48 | $48 |
| Setup labor | $1,634 | $4,085 |
| Total/month | $457 | $1,141 |
| Self hosted, 36 months | $18,092 | $45,156 |
| Alternative total, savings and break even | Quote required | Quote required |
Setup = hours × $68.09. Monthly = servers + (stored GB minus 10) × $0.00695 + administration hours × $68.09. Total = setup + 36 × monthly; savings = alternative total minus self hosted total. Calculations use unrounded values; displayed totals round to whole dollars.
The vendor FAQ requires upgrading to Standard or Custom after the trial. No 36 month Starter retention is assumed. Standard starts at $600/month, so its 36 month advertised floor is $21,600 before usage and setup. That production service is not equivalent to this trial host; savings and break even require matching designs and quotes.
| Zeppelin pilot versus reporting seats | Low case | High case |
|---|---:|---:|
| Our assumption: setup hours | 4 | 12 |
| Our assumption: administration hours/month | 1 | 3 |
| Our assumption: stored backup gb | 100 | 500 |
| Servers/month | $48 | $48 |
| Setup labor | $272 | $817 |
| Total/month | $117 | $256 |
| Self hosted, 36 months | $4,474 | $10,021 |
| Power BI Pro, monthly | $70 | $140 |
| Alternative, 36 months | $2,520 | $5,040 |
| Savings from self hosting, 36 months | Negative $1,954 | Negative $4,981 |
| Break even month | Never | Never |
Setup = hours × $68.09. Monthly = servers + (stored GB minus 10) × $0.00695 + administration hours × $68.09. Total = setup + 36 × monthly; savings = alternative total minus self hosted total. Calculations use unrounded values; displayed totals round to whole dollars.
Our assumption: five to ten paid users and zero separately charged alternative setup. Power BI is a narrower report sharing alternative; it is cheaper in this calculation. Preserve notebook files, interpreter configuration and credentials securely, and test the restore before relying on a result.
### Setup time when Claude or ChatGPT does the typing
We keep the Superset example's one $48 server and 100 to 500 GB of backups. We apply our IT assistance assumptions to its existing IT labor budget, despite using one host. These paired low and high cases are assumptions, not measured installations; human review is included. [Coding studies](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/) do not validate deployment savings.
| Superset: assumptions and 36 month costs | Low case | High case |
| --- | ---: | ---: |
| Manual setup hours | 24 | 60 |
| AI setup share; hours | 65%; 15.6 | 90%; 54 |
| Human review hours within AI setup | 6 | 18 |
| Manual admin hours/month | 6 | 16 |
| AI admin share; hours/month | 80%; 4.8 | 100%; 16 |
| Human review hours within AI admin/month | 3 | 8 |
| Manual total, 36 months | $18,092 | $45,156 |
| Claude Max 5x subscription total, 36 months | $16,479 | $46,647 |
| ChatGPT Pro 100 subscription total, 36 months | $16,479 | $46,647 |
| API: Claude Opus 5.5 total, 36 months | $14,651 | $45,209 |
| API: GPT-6 Astra total, 36 months | $14,766 | $45,902 |
| Assumed setup API input / output tokens | 1,000,000 / 100,000 | 5,000,000 / 500,000 |
| Assumed monthly API input / output tokens | 500,000 / 50,000 | 2,000,000 / 200,000 |
| Opus input / short write / read / output rates per million | $4 / $5 / $0.20 / $20 | $4 / $5 / $0.20 / $20 |
| Astra input / short write / read / output rates per million | $10 / $12.50 / $1 / $50 | $10 / $12.50 / $1 / $50 |
With L = $68.09/hour, S/A = manual setup/monthly hours, q/r = AI shares, H = hosting plus backups, and T0/Tm = setup/monthly tools: manual total = S × L + 36 × (H + A × L); AI total = S × q × L + T0 + 36 × (H + A × r × L + Tm). Round totals after calculation.
[Claude](https://claude.com/pricing) and [ChatGPT](https://chatgpt.com/pricing/) each cost $100 for the selected monthly plan: one setup month, then $50/month (50%) for 36 operating months if other work uses the rest. Dedicated use adds $1,800 to each total. Allowances differ; check limits and budget extra usage. High cases save no admin hours; subscriptions exceed manual costs from month 7.
[API billing](https://developers.openai.com/api/docs/pricing) is separate. Token cost = (ordinary input × rate + cache writes × rate + cache reads × rate + output × rate) / 1,000,000. Our cumulative budgets include retries and billed reasoning: low input is 30% ordinary, 10% short writes, 60% reads; high input is uncached. Use standard rates within context limits.
Claude Code or [Codex](https://developers.openai.com/codex/cli) can draft Superset deployment settings and metadata backup scripts. Desktop Commander can install dependencies, edit Superset configuration and read worker logs through authorized server access; we budget its free local server, excluding optional paid Remote. Claude in Chrome can inspect provider settings, Cloudflare DNS and Superset viewer permissions; ChatGPT Work offers a separate cloud browser, not an assumed Chrome extension.
Use a new or staging server, keep secrets out of chat, and approve commands or restrict them in a sandbox. Watch dashboard actions; review changes, test viewer permissions and restore the metadata database. Use subscriptions through supported supervised tools and API keys for custom automation; check applicable terms and never share account credentials. We still own backups, updates and access decisions.
Our conclusion stays the same: Preset Professional remains cheaper at $4,500 to $9,000 over 36 months; AI assistance does not pay for dashboard design.
Return to the [full Superset comparison](https://jwatte.com/blog/apache-dashboards-analytics-small-business/#setup-time-when-claude-or-chatgpt-does-the-typing).
## Fact-check notes and sources
Official pages below were read October 29, 2026. Catalog rows and source entries without an individual time use the shared directory research from 21:00 to 21:27 UTC; additional IT research ran from 21:37 to 21:42 UTC. Later checks are recorded individually. Registry dates can lag; file dates identify distribution timestamps, not release announcements. We read documentation and pricing; we did not install these products.
1. [superset.apache.org, faq](https://superset.apache.org/docs/faq/), read October 29, 2026, 21:39 UTC.
2. [superset.apache.org, docker compose](https://superset.apache.org/docs/installation/docker-compose/), read October 29, 2026, 21:39 UTC.
3. [superset.apache.org, kubernetes](https://superset.apache.org/docs/installation/kubernetes/), read October 29, 2026, 21:39 UTC.
4. [Index of /superset](https://downloads.apache.org/superset/), read October 29, 2026, 21:38 UTC.
5. [Pricing, Preset](https://www.preset.io/pricing/), read October 29, 2026, 21:38 UTC.
6. [Local quickstart, Apache® Druid](https://druid.apache.org/docs/latest/tutorials/), read October 29, 2026, 21:47 UTC.
7. [Single server deployment, Apache® Druid](https://druid.apache.org/docs/latest/operations/single-server/), read October 29, 2026, 21:47 UTC.
8. [Download, Apache® Druid](https://druid.apache.org/downloads/), read October 29, 2026, 21:38 UTC.
9. [Imply Database as a Service Pricing](https://www.imply.io/pricing/), read October 29, 2026, 21:47 UTC.
10. [Apache Zeppelin 0.12.0 Documentation: Install](https://zeppelin.apache.org/docs/latest/quickstart/install.html), read October 29, 2026, 21:38 UTC.
11. [Download](https://zeppelin.apache.org/download.html), read October 29, 2026, 21:38 UTC.
12. [Power BI: Pricing Plan, Microsoft Power Platform](https://www.microsoft.com/en-us/power-platform/products/power-bi/pricing), read October 29, 2026, 21:39 UTC.
13. [Platform Requirements, Apache Cloudberry (Incubating)](https://cloudberry.apache.org/docs/deployment/platform-requirements/), read 2026-10-29, shared series research.
14. [Hardware and Software Environment Check, Apache Doris](https://doris.apache.org/docs/4.x/install/preparation/env-checking/), read October 29, 2026, 21:39 UTC.
15. [Pricing, VeloDB](https://www.velodb.io/pricing), read October 29, 2026, 21:39 UTC.
16. [Prerequisite, Apache Kylin](https://kylin.apache.org/docs/deployment/prerequisite/), read October 29, 2026, 21:39 UTC.
17. [Single Node Mode, Apache Kylin](https://kylin.apache.org/docs/deployment/single_node_mode/), read October 29, 2026, 21:39 UTC.
18. [Deployment, Apache Pinot Docs](https://docs.pinot.apache.org/operate-pinot/deployment), read October 29, 2026, 21:42 UTC.
19. [BigQuery, Google Cloud](https://cloud.google.com/bigquery/pricing), read October 29, 2026, 21:39 UTC.
20. [Incubation Policy](https://incubator.apache.org/policy/incubation.html), read October 29, 2026, 21:37 UTC.
21. [Apache Cloudberry (Incubating)](https://cloudberry.incubator.apache.org/), read 2026-10-29, shared series research.
22. [DataSketches |](https://datasketches.apache.org), read 2026-10-29, shared series research.
23. [Apache Doris: Open Source, Real-Time Analytics and Search Database for the AI Era](https://doris.apache.org/), read 2026-10-29, shared series research.
24. [Apache Druid, Apache® Druid](https://druid.apache.org/), read 2026-10-29, shared series research.
25. [Apache ECharts](https://echarts.apache.org), read 2026-10-29, shared series research.
26. [Apache Flagon](https://flagon.apache.org), read 2026-10-29, shared series research.
27. [Welcome to Apache Kylin, Apache Kylin](https://kylin.apache.org), read 2026-10-29, shared series research.
28. [Home, Apache Ossie (incubating)](https://ossie.incubator.apache.org/), read 2026-10-29, shared series research.
29. [Apache Pinot, Real-Time Analytics for Users and AI Agents](https://pinot.apache.org), read 2026-10-29, shared series research.
30. [Welcome, Superset](https://superset.apache.org), read 2026-10-29, shared series research.
31. [Zeppelin](https://zeppelin.apache.org/), read 2026-10-29, shared series research.
32. [projects.apache.org, projects.json](https://projects.apache.org/json/foundation/projects.json), read 2026-10-29T21:00:00Z, shared directory snapshot.
33. [Droplet Pricing, DigitalOcean](https://www.digitalocean.com/pricing/droplets), read 2026-10-29, shared series research.
34. [B2 Cloud Storage Pricing, Backblaze](https://www.backblaze.com/cloud-storage/pricing), read 2026-10-29, shared series research.
35. [Occupational Employment and Wage Statistics Profiles](https://data.bls.gov/oesprofile/?areas=INDUSTRY,STATE,MSA&major_group=150000&measure=01&occupation=151244&year=2025), read 2026-10-29, shared series research.
36. [Bureau of Labor Statistics Data](https://data.bls.gov/timeseries/CMU2020000000000P), read 2026-10-29, shared series research.
37. [Claude plan pricing](https://claude.com/pricing), read 2026-10-29T23:31:08.854Z; supplied AI setup research.
38. [Claude Max plan](https://support.claude.com/en/articles/11049741-what-is-the-max-plan), read 2026-10-29T23:31:40.711Z; supplied AI setup research.
39. [Desktop Commander pricing](https://desktopcommander.app/pricing/), read 2026-10-29T23:31:08.785Z; supplied AI setup research.
40. [Desktop Commander setup and permissions](https://raw.githubusercontent.com/wonderwhy-er/DesktopCommanderMCP/main/README.md), read 2026-10-29T23:31:41.305Z; supplied AI setup research.
41. [Claude in Chrome setup](https://support.claude.com/en/articles/12012173-get-started-with-claude-in-chrome), read 2026-10-29T23:31:11.010Z; supplied AI setup research.
42. [Claude in Chrome permissions](https://support.claude.com/en/articles/12902446-claude-in-chrome-permissions-guide), read 2026-10-29T23:31:09.885Z; supplied AI setup research.
43. [Anthropic Consumer Terms](https://www.anthropic.com/legal/consumer-terms), read 2026-10-29T23:31:08.765Z; supplied AI setup research.
44. [METR study of experienced developers](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/), read 2026-10-29T23:30:58.118Z; supplied AI setup research.
45. [OpenAI prompt caching](https://developers.openai.com/api/docs/guides/prompt-caching), read 2026-10-29T23:32:00.479Z; supplied AI setup research.
46. [OpenAI API pricing](https://developers.openai.com/api/docs/pricing), read 2026-10-29T23:31:13.960Z; supplied AI setup research.
47. [Codex CLI](https://developers.openai.com/codex/cli), read 2026-10-29T23:32:01.197Z; supplied AI setup research.
48. [Codex with a ChatGPT plan](https://help.openai.com/en/articles/11369540-using-codex-with-your-chatgpt-plan), read 2026-10-29T23:31:16.001Z; supplied AI setup research.
49. [ChatGPT Work cloud browser](https://help.openai.com/en/articles/20001280-using-cloud-browser-in-chatgpt), read 2026-10-29T23:32:01.431Z; supplied AI setup research.
50. [Anthropic API pricing](https://platform.claude.com/docs/en/about-claude/pricing), read 2026-10-29T23:31:14.425Z; supplied AI setup research.
51. [ChatGPT plan pricing](https://chatgpt.com/pricing/), read 2026-10-29 about 23:27 UTC; supplied AI setup research.
52. [OpenAI Terms of Use](https://openai.com/policies/terms-of-use/), read 2026-10-29 about 23:27 UTC; supplied AI setup research.
*Apache, the Apache feather logo and the Apache project names are trademarks of The Apache Software Foundation. This site is not affiliated with or endorsed by the ASF. Other product names are used for identification only.*