The Roman Map of Britain Bannatia Balnageith?, Moray

Rauatonium (R&C 209) next
Bannatia Bannatia var. Banatia Banatia (Ptolemy II 3 8-9) a polis of the Vacomagi

    Balnageith (NJ0257) and nearby Balnaferry (NJ0357) are recorded on the Pont map as Bannagyith and Banaferry. Banna- combined both with -geith (Gaelic gaoth, earlier gáeth, góith 'a marsh') and with -ferry would suggest that Banna- is a water feature. The River Bann of Dublin derives its name from the goddess Banna. The name of Forres Burn, is obviously a back-formation. It seems the likely candidate for the name Banna. Ptolemy's coordinates indicate that Bannatia is 3° 15' west (south) of Hiberna Pinnatis. The recorded position of Ptolemy's Bannatia could be an error of kd˘ for kz˘d˘˘ (24° for 27° 15' by eye-skip). This would place it near the Moray Firth west of Hiberna Pinnatis.

Blaeu Atlas of Scotland 1645 p.101 (MORAVIA vulgò MURRAY EX CAMDENO - MORAVIA, in the vernacular MORAY. FROM CAMDEN)

Interius ubi iam Bean castrum (Banatia creditur quam memorat Ptolemĉus) vas marmoreum artificiose insculptum & Romanis numismatibus repletum anno MCCCLX fuit repertum. 

Inland, where now is Castle Bean (believed to be the Banatia mentioned by Ptolemy), a marble vase, artistically engraved and full of Roman coins, was found in 1460.

For the record, Camden's Castle Bean seems likely to be Torvean Iron Age fort NH6443 G. torr Beathain.

I've seen scattered about, and without attribution, the identification of Bannatia with the Roman fort at Bochastle NN6107. This does have a potential relationship with Loch Venachar, and I would suspect that is the original basis of the identification.

 James B. Johnston, Place-names of Scotland (1934), p. 320
    Vennachar, L. (Callander). c. 1375 Fordun Banquhar, McFarlane Bennachar, and so = Banchory. But 1794 O. Stat. Acct. says Vennachoir, G. bhanna choir, 'fair corrie' or 'valley.'

    Banchory Devenick and Ternan (Deeside). a. 1300 Banchery defnyk, 1361 -chory deveny; a. 1300 Bancheritarny, 1481 -quhori-terne; also 1164 Benchorin. G. beannachar, 'place of peaks,' beannach, 'peaked, pinnacled'; same as Bangor, Wales (c. 1120 Banchor) and Ireland, Ir. bennchor, 'row of points, circle of peaks,' W. bangor, 'an upper row of rods, a coping, battlement.' Its L. adj. is in Ulst. Ann. 671 Maelrubha Benchorensis. Cf. Beannacher, 1603 Bencar, Kingussie, and L. Beannachar, Ross-sh. 1479 Benquharene, 1508 Kinloch Banquhar, 1571 Beancharan; also 1471 Excheq. R. Tulibanquhara, Perthsh. The -y in Banchory is the old loc. Vennachar is the same. Devenick is fr. St. Devinicus, prob. contemporary of Columba, who worked in Caithness, perh. seen also in Landewedinack, The Lizard. St. Ternan, a. 500, was a Pict. convert of Ninian, then abbot of Bangor, Co. Down.

    Bánnachra (Balloch). 1489 -achar. Prob. = Banchory and Vennachar.

Rivet, PNRB pp126-8, placed Bannatia  not too far away at Dalginross, Perthshire NN7721.